This compilation and all annotations are copyright © Jean Renard Ward, 1996, 1999, 2002, 2005, 2006
Copyright © Fri Sep 5 20:13:49 EDT 2008
Contact: jrward@alum.mit.edu

Mirrors of this page can be found at:

• http://users.erols.com/rwservices/pens/biblio85.html
• http://65.78.28.198:81/pens/biblio85.html
• http://rwservices.no-ip.info:81/pens/biblio85.html

      Main Reference Page      

• [Ahmed84] () Ahmed, P. and Suen, C.Y. "Edge classification and extraction of shape features", Proceedings of the 7th International Conference on Pattern Recognition, 1984, pp 593-596.

  • Li89

• [Aleksander84] (*p) Aleksander, I., Stohnam, T.J., and Wilkie, B.A. "Recognition Apparatus", United States Patent 4,490,847, December 25, 1984, assigned to National Research Development Corporation, London, England.

  • Video image recognition, based on learning (neural) networks

• [ANSI84] () ANSI "Draft Proposed American National Standard Graphical Kernal System", American National Standards Institute, Incorporated, X3.123-198x, 1984.

• [Arcelli85] (*p) Arcelli, C. and Sanniti di Baja, G. "A Width-Independent Fast Thinning Algorithm", IEEE Transactions on Pattern Analysis and Machine Intelligence", Vol PAMI-7 No 4, July 1985, pp 463-474.

  • Skeletonization/line-thinning with a method to "recover" the original outline by building the skeleton back up
  • Line thinning by topology-preserving removal operations

• [Artwick84] (*p) Artwick, Bruce "Applied Concepts in Microcomputer Graphics", Prentice-Hall, 1984, ISBN 0-13-580226-1

  • Describes viewports and Windows as projected coordinate calibration on a digitizing tablet. Early CAD menus printed statically on the digitizer surface as a template.

• [Asija85] (*) Asija, Satya P. "Automated Image Input, Storage and Output System", United States Patent 4,523,331, June 11, 1985

  • Automated image input, including handwriting recognition, and output, including generation
  • Shows characters normalized to a 5x7 grid, and stored based on bit string of traversal of squares
  • Recognition trained to each user: matches are by binary-value distance between two bitstrings
  • Note: Compare with input technolgy of Casio PF-8000 calculator: Casio84_PF8000a

• [Autodesk84] () Autodesk "AutoCAD 2 (tm) Drafting Package User Guide", #106-006, Autodesk Incorporated, 2320 Marinship Way, Sausalito CA 94965, 1984.

• [Baird84] (*p) Baird, H.S. "Model-based Image Matching Using Location", PhD Thesis, Department of Electrical Engineering and Computer Science, Princeton University, October 1984.

  • Pattern matching by a template in feature-space, looking for minimum distances of feature points in a multi-dimensional vector.
  • Mathematical pattern-matching thesis, without any examples of any specific patterns it was tried on!

• [BarkerPG84] () Barker, P.G., Najah, M. and Roper, J.S. "User Experiences with a MICROPAD", Journal of Microcomputer Applications, Vol 7, 1984, pp 19-39.

  • BarkerPG85

• [BarkerPG85a] () Barker, P.G. and Janah, M. "Implementing pictorial interfaces using a high-resolution digitizer", International Journal of Man-Machine Studies, Vol 23, 1985, pp 153-173.

  • User interface picking menus from parts of pictures -- happened to use a Micropad digitizer, not really handwriting recognition

• [BarkerPG85b] () Barker, P.G. and Janah, M. "Pictorial interfaces to data bases", International Journal of Man-Machine Studies, Vol 23, 1985, pp 423-442.

  • User -interface using menus on picture parts, not lists, pick-boxes from a Micropad terminal

• [BeckerJ84] (*p) Becker, Joseph D. "Multilingual Word Processing", Scientific American, Vol 251 No 1, July 1984, pp 96-107.

  • Character/word order, user-interface design for multiple language word processing (Arabic/Hebrew/English/Roman/Chinese)
  • Korean is written in syllable clusters, Hebrew right-to-left, Arabic right-to-left with words diagonally, Mongolian vertically

• [BeckerJ85] (*p) Becker, Joseph D. "Typing Chinese, Japanese, and Korean", IEEE Computer, January 1985

• [Belaid84] (*p) Belaid, Abdelwaheb and Haton, Jean-Paul "A Syntactic Approach for Handwritten Mathematical Formula Recognition", IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol PAMI-6 No 1, January 1984, pp 105-111.

  • Refers to ambiguity of nominally straight vs curved lines
  • Mentions P/D, T/+ ambiguity
  • Investigate high-level context and ambiguity in mathematical formulas
  • Refers to min/max extents, centroid of characters
  • Says some irresolvable "confusion" ambiguity solved by user interaction
  • User interface for 2-D mathematical symbol input: handwritten sketch input

• [Berkin84] (*p) Berkin, G.M., and Negin, Michael "Pattern Recognition System", United States Patent 4,441,205, April 3, 1984, assigned to Kulicke and Soffa Industries, Incorporated, Horsham, Pennsylvania.

  • Alignment pattern recognition for video scanning of integrated circuits, small parts

• [Blesser84] () Blesser, Barry "Comments on Variability due to Pen-Artifacts", internal report, Pencept, Incorporated September, 1984.

  • Apparent variabiity of handwriting in on-line recognition caused in part by defects/performance of tablet digitizers: retrace, hooks, etc.

• [Blesser85] (*) Blesser, Barry "Tablet Error Analysis", internal report, Pencept, Incorporated September, 1985.

  • Mathematical measurements of tablet performance, focusing on defects and quirks of different types of digitizing tablets: describes effects of search algorithms used to locate stylus, local streatch, external voltages, electronic defects, etc.

• [Boies85] (*p) Boies, Stephen J., Gould, John D., Levy, Stephen, Richards, John T., Schoonard, Jim "The 1984 Olympic Message System -- a Case Study in System Design", IBM Research Division Report RC 11138 (#50065), 5/2/85, 53 pages

• [Boivie84a] (*p) Boivie, Richard H. "Directory look-up method and apparatus", United States Patent 4,453,217, assigned to AT&T Bell Laboratories, Murray Hill, New Jersey, June 5, 1984.

  • Spelling correction by recursive division of the dictionary: skip one character at a time: interesting
  • Dictionary searching method

• [Boivie84b] () Boivie, R.H. United States Patent 4,000,000, June 5, 1984, assigned to AT&T Bell Laboratories, Murray Hill, New Jersey.

  • Cited in Boivie85: on pattern matching of chain-code strings
  • Thinning handwritten characters and comparing with dictionary for recognition.

• [Boivie85] (*p) Boivie, R.H. "Character Recognition Arrangement", United States Patent 4,525,860, June 25, 1985, assigned to AT&T Bell Laboratories, Murray Hill, New Jersey.

  • Described curve tracer/follower for OCR of handwritten characters
  • OCR using line thinning, then simulated chain codes generated by a curve tracer/follower: claims to be font-independent / handwriting recognition
  • "standard spelling": coded sequence of chain codes
  • Refers to two approaches for character recognition: theoretic (or discriminant) versus syntactic (or structural) pattern recognition approach: dislikes both
  • Sub-patterns (features) of a handwritten character can only be recognized in the (graphical) context of the entire character pattern

• [Bono85a] (*p) Bono, Peter R. "A Survey of Graphics Standards and Their Role in Information Interchange", IEEE Computer, Vol 18 No 10, October 1985, pp 63-75.

  • User-interface management system (UIMS) and graphics standards

• [Bono85b] (*p) Bono, Peter R. "Software Standards: Which Ones Are Here to Stay?", S. Klein Computer Graphics Review, Inaugural Issue, 1985.

• [Bozinovic84] (*p) Bozinovic, R. and Srihari, S.N. "Knowledge-based Cursive Script Interpretation", Proceedings of the 7th International Conference on Pattern Recognition, Montreal, 1984, pp 774-776.

  • Cites own poor performance if did not exclude "hill and dale" sloppy writing, crowded writing
  • Assumes a premise that people use script, not printing, in normal writing

• [Brault84] (*p) Brault, J.J. and Plamondon, R. "Histogram Classifier for Characterization of Handwritten Signature Dynamic", Proceedings of the 7th International Conference on Pattern Recognition, Montreal, 1984, pp 619-622.

  • Signature verification: small sample size: 243 signatures of 50 people
  • Signature verification: test protocol had no forgery attempts (!)

• [Brocklehurst85] () Brocklehurst, E. "Computer methods of signature verification", Journal of Forensic Science, Vol 30, 1985, pp 445-457.

  • Cited in Yoshimura91

• [Brogardh84] (*p) Brogardh, Torgny "Device for Simultaneous Creation and Detection of Markings", United States Patent 4,475,240, October 2, 1984, assigned to ASEA Aktiebolag, Vasteras, Sweden.

  • Digitizer tablet using a light-sensitive optical sensor to detect markings as they are written/drawn

• [Brogardh85] (*) Brogardh, Torgny "Optical Digitizer/Position Measuring Device", United States Patent 4,531,230, July 23, 1985, assigned to ASEA Aktiebolag, Vasteras, Sweden.

  • Fiber-optic optical digitizer tablet using a grid of an optically-sensible pattern

• [Brooks85] (*p) Brooks, C.P. and Newell, A.F. "Computer Transcription of Handwritten Shorthand as an Aid for the Deaf -- a Feasibility Study", International Journal of Man-Machine Studies, Vol 23, 1985, pp 45-60.

  • Character recognition for shorthand
  • DCR for shorthand much more complex than was thought
  • 1/4 to 1/2 of words in spoken conversation unintelligible out of context
  • Hooks, loops, circles are "real" features in shorthand
  • Refers to endpoint trimming on digitizer strokes
  • "invariant" base form, and numerous real written forms for one writer
  • Hard to write vertical curved strokes reliably

• [BrownE85] (*) Brown, E., Buxton, W., and Murtagh, K. "Windows on Tablets as a Means of Achieving Virtual Input Devices", in D. Diaper et al (Editors), "Human-Computer Interaction - INTERACT '90", Elsevier Science Publishers, pp. 675-681

  • See also http://www.billbuxton.com/windows.html
  • Refers to dynamic windows, virtual devices on tablets, tools for defining regions flexibly. Use touch tablet to simulate keypads, mouse, other devices.

• [BrownG85] (*p) Brown, G.P., et al "Program Visualization: Graphical Support for Software Development", IEEE Computer, Vol 18 No 8, August 1985, pp 27-35.

  • Interactive graphical input and display interface languages for programming
  • Digitizer using separately mounted Lexan sheet (with shift in position?)

• [BrownMK83] (*p) Brown, M.K. and Ganapathy, G "Preprocessing Techniques for Cursive Script Word Recognition", Pattern Recognition, Vol 16 No 5, 1983, pp 447-458

  • adaptive/trainable recognition, suggests "closed loop" verification (feedback to user) is helpful

• [BrownRM85] (*) Brown, Robert M. "Handprinted symbol recognition system", United States Patent 4,491,960, January 1, 1985, assigned to the United States of America as represented by the Secretary of the Navy.

  • Cited in Sklarew88b: on OCR of handwriting?
  • uses logic tree of robust features first, then lesser disciminants, then a quality-assuance module
  • filtering/noise-removal: (1) pre-recognition ("trash" filter), (2) potential symbol identification, and (3) final quality assurance; for on-line handwriting recognition

• [Brugge92] (*) Brugge, John "Introduction to PenPoint", MacTECH,The Journal of Macintosh Technology: available at http://www.mactech.com/articles/frameworks/6_2/PenPoint_Brugge.html

  • Overview article from about one year after PenPoint OS announced

• [BuxtonW85] (*p) Buxton, William "Notes on Chord Keyboards", Computer Systems Research Institute, Univ. of toronto, Toronto Ontario CA M5S 1A4

  • mentions WriteHander chord keyboard from NewO Corporation

• [BuxtonW85a] (*) Buxton, William, Hill, Ralph, and Rowley, Peter "Issues and Techniques in Touch-Sensitive Tablet Input", SIGGRAPH 85, San Francisco Jly 22-26, Vold 19 No 3, 1985, ACM 0-89791-166-0/85/007/0215, also Computer Graphics Vol 19 No 3

  • Also available at http://www.billbuxton.com/touch.html
  • Describes virtual tablets, multiple virtual devices on a tablet. Use of pressure/force and multiple-touch input devices.

• [BuxtonW85b] (*) Buxton, W., Sniderman, R., Reeves, W., Patel, S. and Baecker, R. "The evolution of the SSSP score-editing tools", In Foundations of Computer Music, C. Roads and J. Strawn, Eds, MIT Press, Cambridge, Mass 1985, ch 22, pp 387-22; Originally published in Computer Music Journal, Vol 3 No 4 Issue 12, December 1979; pp 14 ff.

  • Gestures/marks for editing musical scores, handwriting music recognition. Recognition based on changes in angles only, not lengths of segments. Semantically uses multiple segments to indicate note duration, but does not treat as compound recognition. Shapes later copied by PLI group at IBM for music editing demo.
  • Also available at http://www.billbuxton.com

• [CalComp85] () CalComp "CalComp 2500 Series Digitizer Operator's Manual User's Guide", CalComp Incorporated 50293-1, 1985.

• [Carnes84] () Carnes, W.R. "Signature Verification System", United States Patent 4,443,436, February 21, 1984.

  • From patent data base

• [Carpenter85] (*p) Carpenter, G.A. and Grossberg, S. "Neural Dynamics of Category Learning and Recognition: Attention, Memory Consolidation, and Amnesia", in "Brain Structure, Learning, and Memory", Davis, J., Newburgh, R., and Wegmen, E. editors, AAAS Symposium Series, 1985.

  • Adaptive: refers to "as recognition categories stabilize" in learning

• [CarrTH84] () Carr, T.H., Brown, T.L., Brown, J.S., and McDonald, J.L. "Handwriting of connected discourse: Speed-accuracy tradeoffs in a skilled performance", Proceedings of Psychonomic Society, November, 1984.

  • Cited in Rhyne86: : was in Tappert's group 1986
  • Rhyne86 cites this that mental set affects legibility of handwriting (variability)

• [Casey84] () Casey, R.G. and Nagy, G. "Decision Tree Design Using a Probabilistic Model", IEEE Transactions on Information Theory, Vol IT-30 No 1, January 1984, pp 93-99.

  • Applications: single-font OCR system had to replace five keypunch operators at prices of 1975
  • Parallelism vs serial: decision tree is best technique for speed/cost tradeoffs in character recognition
  • Paradigm: iterative optimization of decision tree by designer vs automatic construction of decision tree (designer specifies features to use)
  • Decision trees in OCR: for small trees, need very uniform styles and pixel data (is not that obvious?)
  • Estimated versus actual performance: adding more OCR pixels does not help much since they are not significant features (size of character/resolution)
  • Testing: reports real results on trading reject for substitution errors

• [Casio84_PF8000Eng] (*) Casio "Casio PF-8000 Calculator (in English and Spanish)", Casio 123B M publication, date believed to be 1984

  • 1984 Calculator with touchscreen input, and also zone-based handwriting recognition on a touch-film keyboard/keypad, separate from the display
  • Shows a "gesture" character for editing (e.g. backspace/delete), demonstrated in a name/address-book application included in the unit.
  • Functional unit in library, this is the user documentation
  • unistroke / single-stroke characters
  • Case includes a reference sheet for writing single-stroke characters, similar to the "Graffiti" unit by U.S. Robotics in the 1990s

• [Casio84_PF8000a] (*) Casio "Casio PF-8000 Calculator (in Japanese)", Collection from g-mark.org, http://www.g-mark.org/library/supercollection/g40/1984/index-20.html, 2002

  • 1984 Calculator with touchscreen input, and also zone-based handwriting recognition on a touch-film keyboard/keypad
  • Not sure about relationship of PF8000 and IF8000 products
  • unistroke / single-stroke characters (?)
  • Dr. Theodore Kuklinski had one in his collection in 1990
  • Additional reference on Casio/Cacio Data Bank PF-8000: http://www.geocities.co.jp/Technopolis/6766/PF-8000.htm, full-frame picture
  • Additional reference on Casio Data Bank PF-8000: http://www.steamtrain.de/floh/bilder_2/casio_data_bank.jpg, showing writing styles

• [Casio84_PF8000b] (*) Casio "Casio PF-8000 Calculator user manual (in Japanese)", In personal collection of Jean Renard Ward -- cover missing, approximately 34 pages

  • 1984 Calculator with touchscreen input, and also zone-based handwriting recognition on a touch-film keyboard/keypad

• [Castleberry85] (*) Casteleberry, Donald E. and Becker, Charles A "Electroscopic information display and entry system with writing stylus", United States Patent 4,520,357 May 28, 1985, assigned to General Electric Company

  • Sklarew cites as electroscopic tablet/display integrated device
  • Capacitive-coupling digitizer integrated with a matrix display

• [Chainer85a] (*p) Chainer, T.J., and Worthington, T.K. "Segmentation Algorithm for Signature Verification", United States Patent 4,553,258, November 12, 1985, assigned to International Business Machines, Armonk, New York.

  • Finds pen lift/down stroke start/end via A.C. coupling to a force/pressure gauge

• [Chainer85b] (*p) Chainer, T.J., Gundersen, S.C., and Worthington, T.K. "Semi-Independent Shifting Technique for Signature Verification", United States Patent 4,553,259, November 12, 1985, assigned to International Business Machines, Armonk, New York.

  • Force/pressure and acceleration information on signature verification correlating data by shifting it

• [Chainer85c] (*p) Chainer, T.J., Scranton, R.A. and Worthington, T.K. "Data Input Pen for Signature Verification", United States Patent 4,513,437, April 23, 1985, assigned to International Business Machines, Armonk, New York.

  • Better force/pressure-sensitive (piezo-electric sensor) pen stylus than Herbst 4,142,175
  • Digitizer stylus pen with force/pressure and acceleration sensitivity for signature verification.

• [Chainer85d] (*p) Chainer, T.J. and Worthington, T.K. "Shifting Technique for Signature Verification", United States Patent 4,562,592, December 31, 1985, assigned to International Business Machines, Armonk, New York.

  • Shifting (to reduce correlation computation) of segments of signatures for acceleration and force/pressure data for signature verification

• [Cheung85] () Cheung, Y.S. and Leung, C.H. "Chain-code transform for Chinese character recognition", Proceedings of IEEE International Conference on Systems, Man, and Cybernetics", Tucson, Arizona, 1985, pp 42-45.

  • Cited in Leung87
  • Japanese/Kanji/Katakana product from Hew Crane and James Dao

• [CIC84] () CIC "Handwriter (TM) Product Literature", CIC Japan Incorporated, Suite 918, Shuwa Kioicho TBR Building, 5-7 Kojimachi, Chiyoda-Ku, Tokyo 102, Japan, 1984.

  • Forms data entry from CIC Handwriter product

• [CIC85a] () CIC "Handwriter (R) focus:ABC (TM) Accounting Package", Communication Intelligence Corporation, 800 El Camino Real, Menlo Park, CA 94025, 1985.

  • Spreadsheet from CIC Handwriter product

• [CIC85b] () CIC "Handwriter (R) Lotus 1-2-3 interface Kit", Communication Intelligence Corporation, 800 El Camino Real, Menlo Park, CA 94025, 1985.

  • Word-processing from CIC Handwriter product

• [CIC85c] () CIC "Handwriter (R) WordStar (R) interface Kit", Communication Intelligence Corporation, 800 El Camino Real, Menlo Park, CA 94025, 1985.

  • The ASCII writing product Hew Crane said does not exist

• [CIC85] (*) CIC "Handwriter (R) GrafText (TM) System Model GT-5000", Communication Intelligence Corporation, 800 El Camino Real, Menlo Park, CA 94025, June 1985.

  Similar product to Pencept PenPad. Describes recognition modes and templates for forms: graphic, text, remot-function, local-function areas. Calibration of tablet to paper form. Handwriting recognition returned multiple matches (up to 5).

• [CItoh85] () CItoh Corp "CI-Handwriter Fit-In-1 product information", 1985. (OEM version of CIC Handwriter product).

  • Japanese handwriting character recognition product

• [Cobblestone85] (*) Cobblestone Design "Pencept: Software Control at the Stroke of your Pen / Write into your computer", www.cobblestonesdesign.com/Pages/frames/Pencept.html

  • Photograph of trade-show booth from Comdex for Pencept, handwriting recognition.

• [Comerford84] () Comerford, Richard "Pointing-device innovations enhance user/machine interface", Electronics Design News, July 26, 1984, pp 54-66.

  • Review of digitizer/tablet/joystick/mouse/touchscreen/trackball vendors, 1984

• [Convis85] () Convis, D.B., Grim, P.J., and Reed, M.A. "Retro-stroke Compression and Image Generation of Script and Graphic Data Employing an Information Processing System", United States Patent 4,550,438, October 29, 1985, assigned to IBM Corporation, Armonk, New York.

  • Uses local-maxima chain codes for image compression of script

• [Cook84] (*) Cook, Mike "Light Pen", The Micro User, Volume 2, Number 1, March 1984

  • Light pen - software techniques (via rubber-banding) to get single-pixel resolution from a light pen that only has resolution to character cells in a character CRT display

• [Crane85a] (*) Crane, H.D., Ostrem, J.S., and Edberg, P.K. "Method for Distinguishing between Complex Character Sets", United States Patent 4,531,231, July 23, 1985, assigned to Communication Intelligence Corporation, Menlo Park, California.

  • Attempt to patent Pencept's "area" definition on Penpad 320: Kanji (Chinese) in one area, Kana/Roman in another (Katakana/Hiragana Japanese)
  • Kanji/Kana by top half / full height of character box (similar to upper/lower case)

• [Crane85b] (*) Crane, H.D., and Ostrem, J.S. "Complex Pattern Recognition Method and System", United States Patent 4,561,105, December 24, 1985, assigned to Communication Intelligence Corporation, Menlo Park, California.

  • Chines/Kanji character recognition using stroke shape and curvature
  • Distance metrics between strokes for handwriting recognition of Chinese/Kanji characters
  • Allows for shift in position, rotation of strokes to allow for better recognition mathces

• [Davies85] () Davies, K. "Continuous Speech Recognition", Probe Research Seminar: The Coming of Age of Voice Recognition, San Francisco California, 1985.

  • Cited in Mangione86: : member of IBM Yorktown Hts. staff (with Tappert)
  • Adaptive recognition dies trying to get complete set of samples

• [DosterW84a] () Doster, W. and Oed, R. "Textbearbeitung auf Personal-Computern mit handschriftlicher Direkteingabe", presentation at Arbeitstagung der Fachgruppe Interaktive Systems der Gesellschaft fuer Informatik, March 1984, in Zurich, Switzerland. Reprinted in PC-Praxis. Author's address: AEG Aktiengesellschaft, Research Center Ulm, Sedanstrasse 10, D-7900 Ulm, West Germany.

  • Points out their variability model cannot be statistically verified
  • Describes AEG's general DCR work-station, including driver design

• [DosterW84b] (*) Doster, W. and Oed, R. "Word Processing with On-line Script Recognition", IEEE Micro, October 1984, pp 36-43.

  • AEG's word-processing user interface for handwriting recognition input. Describes word processing (for handwriting user interface) as two phased: text creation, and text editing. Some use of gestures or recognition macros, separate gesture, pointing (virtual tablet) and text input areas on separate digitizer from display. Compare with PenCept PenPAD and recognition macros, virtual tablet definitions.

• [DosterW84c] () Doster, W. "Das papierlose Buero: Lesemaschine fuer Handschriften", Funkschau (West Germany), Vol 25, 1984, pp 61-63.

  • General article on AEG's handwriting recognition system for forms data entry and word processing in an electronic automated office

• [DosterW85a] () Doster, W. "Pattern Recognition Techniques as a Stimulus for Advanced Text Processing", PROTEXT II: Proceedings of the 2nd International Conference on Text Processing Systems, October 23-35, Dublin, Ireland, pp 155-161.

  • Says putting in multiple variants a deficiency of non-adaptive
  • Says "user-definable" special gesture symbols (for WordStar word processor) better for command, but his examples are of modified standard characters
  • Points out that non-adaptive systems are user-adaptive, and results are the same "perfect" mutual enrollment

• [DosterW85b] () Doster, W. and Schuermann, J. "Bildanalyse von Textdokumenten und Handschriftliche Direkteingabe - Zwei Verfahren Auch fuer BTX-Anwendungen", Informatik Fachberichte 92, Springer-Verlag, New York, 1985, pp 247-265.

  • Two research projects at AEG on the automated electronic office: automatic document analysis for O.C.R. images of text and pictures, and on-line handwriting recognition

• [Downton83] () Downton, A.C., Baker, R.G., Lewis, S.M., and Cooper, P.J. "Readability measurements of Palantype transcription for the Deaf", International Journal of Man-Machine Studies, Vol. 19, pp 575-594, 1983

• [Downton84] () Downton, A.C. and Brooks, C.P. "Automated machine shorthand transcription in commercial applications", Proceedings of INTERACT '84, IFIP Working Group 6.3, London, United Kingdom, September 4-7, 1984, pp 151-156.

  • Need whole conference proceedings: user interface with handwriting editing.

• [Dye84] () Dye, R., Newell, A.F., and Arnott, J.L. "An adaptive editor for shorthand transcription systems", Proceedings of INTERACT '84, IFIP Working Group 6.3, London, United Kingdom, September 4-7, 1984, pp 157-162.

  • Need whole conference proceedings: user interface with handwriting editing.

• [Ejiri85] (*) Ejiri, Koichi "Method of Assigning Direction Code to Boundary Pictur eElement in Character Recognition System", United States Patent 4.524,454

  • Line-thinking, extraction of plausible writing direction from edges of scanned OCR images of characters, then to encode as direction codes / chain codes
  • (There is another paper on this sort of thing in here someplace ...)

• [Ewing85] () Ewing, J., Mehrabanzad, S., Sheck, S., Ostroff, D. and Shneiderman, B. "An Experimental Comparison of a Mouse and Arrow-Jump Keys for an Interactive Encyclopedia", to appear in International Journal of Man-Machine Studies, September, 1985. Author's address: Department of Computer Science, University of Maryland, College Park, Maryland 20742.

  • Pre-publication version only
  • User-interface on pointing methods: applies to mouse emulation on a tablet?

• [Flurry85] () Flurry, G.A. "Electronic Handwriting Facility", IBM Technical Disclosure Bulletin, Vol 27 No 9, February 1985, pp 5364-5366.

  • Description of IBM's handwriting work-station?

• [FoleyJD84] () Foley, J.D., Wallace, V.L., and Chan, P. "The Human Factors of Computer Graphics Interaction Techniques", IEEE Computer Graphics and Applications, Vol 4 No 11, November 1984.

  • The granddaddy of all user-interface papers for interactive graphics

• [FormsAutomation85] () Forms magazine "One-step data entry system links forms with computers", forms automation product review, April, 1985, pp 34-41. Published by National Business Forms Association, 443 East Monroe Avenue, Alexandria VA 22301.

  • Description of DataPad/ AnnoGraphics / Portable Computer handwriting recognition product: was marketed under several different name.

• [Fox84] () Fox, A.S., Kim, J. and Tappert, C.C. "Segmenter for Known Number of Characters", IBM Technical Disclosure Bulletin, Vol 27 No 7A, December 1984, pp 3691-3693.

  • Segmentation when writing prototypes, because number of shapes is known

• [Gaines84] () Gaines, B.R., McKellar, I.D., Dinger, W.P., and Fast, S.R. "Some Experience in the Real-time Processing of Handwriting", Proceedings of the 7th International Pattern Recognition Conference, Montreal, 1984, pp 630-632.

  • Says natural user-interface is speech and handwriting together
  • Says tablet digitizers with 1000x1000 resolution and 1% linearity good enough for handwriting (!)
  • Say peak handwriting speed/velocity is 50 inches/second (?), normal is 2.5"/sec
  • Says "skilled" user gets 97%-100%, "unskilled" only 70% recognition success

• [Genter85] () Genter, R.E. "Guidelines for C Program Development", Internal report, Intermetrics, Incorporated, Cambridge, Massachusetts, September 1985.

  • Programming/coding style

• [Gharachorloo85] (*) Gharachorloo, N. "On-line Character Recognition Using Closed-Loop Detector", United States Patent 4,495,646, January 22, 1985.

  • Tablet-free stylus using mechanical design to detect rough stylus direction of motion
  • Claims handwriting recognition using mechanical sensing of chain codes

• [Glenn84] () Glenn, W.E. "Apparatus for Determining Position and Writing Pressure", United States Patent 4,488,000, December 11, 1984, assigned to New York Institute of Technology, Old Westbury, New York.

  • Force/pressure-sensitive digitizer tablet using acoustic sound waves in a sheet

• [Goldwasser85] (*) Goldwasser, Eric, and Goldwasser, Dorothy "Method of Creating Txt Using a Computer", United States Patent 4,559,598, December 17 1985

  • Pen/stylus or other pointing means, select words from dynamic menus to input text. Intended as aid to the blind: compare with Qwikwriting and other means for one-handed input of handwritten text with contrained recognition.

• [Gould84] () Gould, J.D. and Alfaro, L. "Revising Documents with Text Editors, Handwriting-Recognition Systems, and Speech-Recognition Systems", Human Factors, Vol 26 No 4, August 1984, pp 391-406.

  • Applications: text editing much faster with handwriting than with keyboard, even though same number of characters input
  • User-interface: differences in external feedback greatly affect operational speed
  • User-interface: time spent inputting a command much less than think time (and typing adds to effective think time)
  • User-interface: electronic ink would eliminate spatial displacement within written copy and screen copy
  • Speech simulation of handwriting editing, but no standard proof-editing marks
  • Speech: no visual feedback a minus on speech recognition test
  • Speech: USERs make more errors in speech than with writing or typing
  • Martin89 cites as saying most of time in producing a letter is think time, so user-interface of speech recognition or typing not that great a difference

• [Graham84] () Graham, Martin "Signature Verification Sensor", United States Patent 4,475,235, October 2, 1984, assigned to Rolm Corporation, Santa Clara, California.

  • Capacitive plate digitizer tablet

• [GSS84] () GSS "GSS-Drivers User's Guide", Second Edition, Graphic Software Systems, Incorporated, March 1984.

• [Hayamizu84] () Hayamizu, S. and Oka, R. "Experimental Studies on the Connected-Words Recognition using Continuous dynamic Programming", Denshi Tsushin Gakkai Ronbunshi, Vol 67-D No 6, June 1984, pp 677-684 (in Japanese), translated in Systems, Computers and Controls, Vol 15 No 4, 1984, pp 80-88.

  • Speech recognition: performance goes down with vocabulary size
  • (everybody should show performance with increasing vocabulary size)
  • Recognition errors are confusion (substitution), missing, and "ghost word"

• [Higgins84] () Higgins, C.A. and Whitrow, R. "On-line Cursive Script Recognition", Proceedings of Interact '84, 1st IFIPS Conference on Human-Computer Interaction, September 1984, pp 140-144.

  • Unlikely that "unconstrained" can be done, since humans find stuff illegible
  • Segmentation at turning points, not Y-maxima/minima
  • Features must have enough variety to be effective, but not be too sensitive
  • Extra feature tests ("optional") for specific characters
  • Context - "higher shape templates" for combined shapes
  • Refers to including new shapes on the fly for mis-recognition
  • Training is "manual: results in inaccurate feature values for templates" (?)
  • From incorrect weighting, best word match does not always come out on top

• [HillR85] . Hill, R. and Rowley, P. "Issues and Techniques in Touch-Sensitive Tablet Input", Computer Graphics, Vol 19, No 3, pp 215-224.

  • Cited in Wanderley00

• [Hitachi84] () Hitachi, Limited: "Humanication product literature", (in Japanese), 1984.

  • Hitachi Kanji/Chinese handwriting recognition product

• [House84] () House "(title)", United States Patent 4,444,998

  • Lukis87 on digitizer tablets

• [Huang85] () Huang, J.K. "The input and output of Chinese and Japanese characters", IEEE Computer, Vol 18, January 1985, pp 18-24.

  • 50,000 total Chinese characters: three-corner code method TCCM for encoding and indexing for keyboard input

• [Hulls85] (*) Hulls, L.R. "Absolute Position Coordinate Determining Device employing a Single Phase Difference Measurement to Determine Cursor Position", United States Patent 4,552,991, assigned to Numonics Corporation, Lansdale Pennsylvania, November 12, 1985.

  • Phase-shifting tablet digitizer (electromagnetic?), using not a simple grid, but in each ordinate a mixed set of loops at physically different phase.

• [IBM85] () Anonymous "Multi-Segment System for Recognizing Cursive Writing", IBM Technical Disclosure Bulletin, Vol 17 No 11, April 1985, pp 6735-6739.

  • Segmentation algorithm for script (Tappert?) for run-on and script handwriting

• [IchikawaS84] () Ichikawa, S. and Gyoba, J. "Methodological Problems on Pattern Psychophysics", (Japanese journal), Vol 27 No 2, 1984, pp 132-157.

  • Not sure of English name of Japanese journal
  • Similar to functional attributes, describes general problem of figuring out human recognition mechanisms for multiple features

• [Ikeuchi84] () Ikeuchi, K. "A Model of Character Recognition by Humans", Proceedings of the 7th International Pattern Recognition Conference, Montreal, 1974, pp 521-524.

  • Three stages of human recognition: feature extractor, naming, decision
  • Human features: straight line, turned line, end point, curved line

• [Jansen84] . Jansen H. et al "Interpretation of freehand drawings for mechanical design processes", Computers and Graphics, Vol 8 No 4, pages 351-369, 1984: Pergamon Press Ltd, USA

• [Jelinek85] () Jelinek, F. "The Development of an Experimental Discrete Dictation Recognizer", Proceedings of IEEE, Vol 73 No 11, November 1985, pp 1616-1624.

  • Speech: gives results of (huge) number of samples needed to get statistically useful vocabulary coverage
  • Speech: interface problem for word processing is not a trivial one
  • Speech: editing (pointing) with pencilfundamentally better user interface than voice/speech recognition
  • Speech: for editing, best interface will include digitizer and handwritten character recognition for editing marks (mode separation? ref. Buxton "tension")
  • Speech: an adaptive system that does worse with non-native speakers of English (!) (Features are biased? Not really a general adaptive system?)
  • Adaptive speech: cites user who did worse when analyzing own pronunciation (ref: problems of user understanding enrollment problems)

• [KADesign85] () KA Design group "business plan", March 1, 1985.

  • Puck/pointer: proposed mouse-like tablet product using a small joystick.
  • Contains patent references to Victor B. Kley, 4,435,616, Graphical Data Entry Apparatus, resistive sheet tablet digitizer

• [Kahan85] () Kahan, S. "Problems in recognizing hand-printed characters", M.S. Thesis, Department of Electrical Engineering and Computer Science, University of California at Berkeley, February 1985.

  • Cited in Kahan87

• [Kai-tungHuang85] () Kai-tung Huang, J. "The Input and Output of Chinese and Japanese Characters", IEEE Computer, Vol 18 No 1, January, 1985, pp 18 - 24.

  • Chinese and Japanese Kanji input via keyboard

• [Karat84] () Karat, J., McDonald, J., and Anderson, M. "A comparison of selection techniques: touch panel, mouse, and keyboard", Tecnical report No TR-51.0166, IBM Corporation, Entry Systems Division, Austin, Texas, 1984.

  • Mack89, user-interface on integrated tablet/display

• [Kasanuki85] . Kasanuki et al "Sophisticated Man-Machine Systems (2)", 30th Conference of the Data Processing Society of Japan, pp 1831-1832, Japan

  • Cited in Sklarew patents

• [Kataoka84] (*) Kataoka, Hiroyuki and Arai, Yoshio "Character Shaping Device", United States Patent 4,440,513 assigned to Fuji Xero Co., Ltd. Tokyo Japan, April 3, 1984

  • Handwriting text editing: Normalizes size, scale, position, shape of handwritten characters prior to printing to be neatly spaced and lined up better, but does not talk about recognition per se

• [KimJ84] () Kim, J. and Tappert, C.C. "Handwriting Recognition Accuracy versus Tablet Resolution and Sampling Rate", IEEE 7th International Conference on Pattern Recognition, 1984, Montreal Canada, pp 917-918.

  • A "nothing" paper: recognition goes down with worse tablet resolution

• [Konneker84] () Konneker, L.K. "A Graphic Interaction Technique Which Uses Gestures", IEEE Proceedings of First International Conference on Office Automation, page 51-55, 1984.

  • Cited in Rhyne86, Buxton87
  • Rhyne86 cites this as an interesting handwriting input system: Also Goodisman91
  • "C" code is not necessarily portable

• [Krieger85] () Krieger, Mark "Writing Portable UNIX Software", talk given at the 1985 UNIFORUM conference in Dallas, Texas by Mark Krieger, January, 1985. Speaker's address: UniPress Software Incorporated, Suite 312, 2025 Lincoln Highway, Edison NJ 08817.

• [Kuklinski84] (*p) Kuklinski, T. "Components of Hand-Print Style Variability", Proceedings of the IEEE 7th International Conference on Pattern Recognition, pp 924-926, 1984.

• [Kuklinski85] (*p) Kuklinski, T. "A Case for Digitizer Tablets", Computer Graphics World, May, 1985, pp 45-52.

• [Kuklinski85b] (*) Kuklinski, Theodore) "Internal speaking notes, Development of Pencept Digitizing Tablet", Pencept, Incorporated, 39 Green Street, Waltham, Massachusetts 02154, 1985.

  • History of digitizer performance and technology problems that lead to development of Pencept electromagnetic digitizing tablet with two-coil sensing stylus to correct for stylus tilt. Credits to Dr. Barry Blesser. Contains discussion of performance problems of digitizing tablets in general.

• [Kurta85a] (*) Kurta Corporation "Equipment/Device: Kurta IS/PenMouse digitizing tablet, ca. 1985", Kurta Corporation

  • Physical device in collection: PenMouse is a digitizing tablet with a mouse-emulation mode. (Stylus missing).

• [Kushnir85] () Kushnir, M., Abe, K., and Matsumoto, K. "Recognition of Handprinted Hebrew Characters using Features Selected in the Hough Transform Space", Pattern Recognition, Vol 18 No 2, 1985, pp 103-114.

  • Features are stroke direction, concavities, endpoints, intersections, etc.
  • Says high dimensionality of feature space means template matching takes up lots of memory and compute time
  • Handwriting recognition requires new strategies beyond printed recognition
  • Accuracy test uses larger training set than test set
  • Statistics fail by grouping similar characters, which must be solved by heuristics (!)
  • Issues of test samples versus training samples
  • Source of errors is similar characters: boundary?

• [Kuzunki85] . Kuzunki, S. et al "Handwritten Characters and Graphics Recognition Technology for Workstations", Hitachi Hyoron, Vol 67 No 3, pp 243-246, 1985

  • Cited in Sklarew patents

• [Lamarche84] () Lamarche, F. and Plamondon, R. "Segmentation and Feature Extraction of Handwritten Signature Patterns", Proceedings of the 7th International Conference on Pattern Recognition, 1984, pp 756-759.

  • Refers to time-axis distortion (variability) in signatures

• [Lamb84] () Lamb, M.R. and Buckley, V. "New techniques for gesture-based dialogue", Proceedings of INTERACT '84, IFIP Working Group 6.3, London, United Kingdom, September 4-7, 1984, pp 135-138.

  • Need whole conference proceedings: user interface with handwriting editing.

• [LeeSK85] (*) Lee, S.K., Buxton, W. and Smith, K.C. "A Multi-Touch Three Dimensional Touch-sensitive Tablet", CHI'85 Proceedings, pp 21-ff, ACM 0-89791-149-0/85/004/0021

  • capacitive digitizer technology, used for sensing multiple touch points of user's finger on surface: uses recursive division to determine position: used to emulated a piano keyboard

• [Leedham84] () Leedham, C.G., Downton, A.C., Brooks, C.P. and Newell, A.F. "On-line acquisition of Pitman's handwritten shorthand as a means of rapid data entry", Proceedings of INTERACT '84, IFIP Working Group 6.3, London, United Kingdom, September 4-7, 1984, pp 145-150.

  • Need whole conference proceedings: user interface with handwriting editing.

• [Lerner8?] (*p) Lerner "Understanding speech proves tough task for machines", (reference not known, date uncertain: 80s?)

• [Linus84] (*p) Linus Technologies "Business Plan for TechBook Limited Partnership", 5265 Port Royal Road, PO Box 1322, Springfield, Virginia, 22151, February 1984.

  • Original business plan for Linus portable handwriting product: also describes gesture-based user interface

• [Loken-Dim85] (*) Loken-Kim, K.H. "Artificial Intelligence Techniques and their Application to the Correction of Automatic Speech Recognition Input", Proceedings of Voice I/O Systems Applications Conference, September 10-12, 1985, San Francisco, pp 425-441.

  • Speech: context correction requires use of linguistic and extra-linguistic knowledge
  • Speech: high recognition rate is not enough: user interface is important
  • Speech: substitution (modal) recognition errors more severe than rejection in performance perception
  • Speech: sequential restrictions in language are called syntax

• [Lukis85] (*p) Lukis, L.J. and Duhig, G.P. "Character Recognition Device", United States Patent 4,493,104, January 8, 1985, assigned to Moore Business Forms, Incorporated, Grand Island, New York.

  • Pressure-sensitive resistive-sheet digitizer
  • Moore Business Forms DCR patent

• [Lyon85] (*p) Lyon, Richard F. "Cursor Control Device", United States Patent 4,521,772 June 4, 1985, assigned to Xerox Corporation, Stamford, Connectituc

  • Optical mouse: optical digitizer using passive printed pattern: but recognizes relative motion only, not absolute position.
  • (See other papers in file for additional paper and Web references)

• [Lyon85a] (*p) Lyon, Richard F. "Imaging Array", United States Patent 4,521,773 June 4, 1985, assigned to Xerox Corporation, Stamford, Connecticut

  • Optical mouse sensor, with auto-correlation in analog (?) circuitry
  • (See other papers in file for additional paper and Web references)

• [Mandler85] (*) Mandler, Egerhard, Oed, Richard, and Doster, Wolfgang "Experiments in On-Line Script Recognition", Proceedings of 4th Scandinavian Conference on Image Analysis, Norwegian Institute of Technology, Trondheim, Norway, June 17-20, 1985.

  • Gives many European writing variation styles for "A": features are pathlength and arc; segmentation into single characters without context, not limited to alphanumeric characters,

• [Meads85a] () Meads, Jon A. "Report on the SIGCHI Workshop on Planning for User Interface Standards", Jon Meads and Associates, 2516 NE 19, Portland, OR 87212, 1985.

• [Meads85b] () Meads, Jon A. "Friendly or Frivolous?", Datamation, April 1, 1985, pp 96-100.

  • User-interface user-friendly: users do NOT plan ahead: Fitt's Law says a mouse may be user-friendly, but not a mouse and a keyboard

• [Meisel85] (*p) Meisel, W.S. "Empirical Artificial Intelligence in Speech Recognition", Proceedings of the 1st Annual Artificial Intelligence and Advanced Computer Technology Conference, Long Beach, California, April 30 - May 2, 1985, pp 182-187.

  • Empirical artificial intelligence: cross between statistical and knowledge-based expert systems analysis
  • Qualitative vs quantitative data/information
  • AI research community has ignored empirical knowledge in expert systems
  • AI: extracting intelligence from empirical evidence vs from human experts
  • AI expert systems: possible flawed assumption is that humans can accurately tell you what they know
  • Curse of dimensionality: your data may just not happen to show a particular feature or interaction among features
  • Sample size vs number of features can be treated formally
  • EAI: using expert knowledge to uncover structures, and using data to verify structures (is not this just normal scientific method?)
  • Reducing "intrinsic dimensionality" of data and features
  • Need off-line (expert?) development of recognition code/tables/features, because too complex to be done in real-time or on-line
  • Human opinion/context has a large role in recognizing/labeling a speech sound
  • Collect speech data for testing from real-world business correspondence and memos
  • linguistic context rules applied to speech recognition

• [Minsky84] (*) Minsky, M.R. "Manipulating simulated objects with real-world gestures using a force and position sensitive screen", Computer Graphics Vol 18 No3, July 1984, pp 195-203.

  • Touch/pressure-sensitive digitizer tablet: uses four strain guages / load cells at corners supporting a tablet to sense both position, and strength of force
  • Gestures include tapping, not just shape recognition: note that Penpad also had a "click" gesture of sorts for mouse/character-recognition mode

• [Mizzi85] (*) Mizzi, Pierre M.A. "Hand-Held Computer", United States Patent 4,545,023, assigned to Engineering Project Development Limited, London, England, October 1, 1985

  • Hand-held "slate" style computer with no keyboard, provision of phoneme and/or handwriting recognition, and for TV/Radio reception via plug-in cards

• [MoriS84] (*p) Mori, S., Yamamoto, K., and Yasuda, M. "Research on Machine Recognition of Hand-printed Characters", IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol PAMI-6 No 4, July 1984, pp 386-405.

  • It's important to know how well DCR system works on low-quality data
  • Recognition of hiragana (small character set) about as difficult as Kanji (large character set)
  • General review of Kanji character recognition
  • Gives figures for total Kanji characters in use in Japan, and in China
  • Mentions earliest commercial Kanji recognition products, ten current
  • Cites a very large data-base collection, which "scope is not sufficient"
  • Cites Blesser
  • Mentions fundamental differences of human vs machine recognition (p 403)
  • Variations unlimited: must be limited for practicality and academic work

• [MoriS85] (*) Mori, Shunji; Yamada, Hiromitsu; Takeuchi, Kunihiko,; and Do, Michihisa "Feature Extraction System for Digitized Character Information", United States Patent 4,468,808, August 28, 1984

  • Feature extraction for optical character recognition: compare to chain codes, sub-segments with concavity or convexity, loops: states that convexity/concavity noise is eliminated automatically by processing

• [Morita85] (*p) Morita, M. "Japanese Text Input System", IEEE Computer, Vol 18 No 5, 1985, pp 29-35.

  • Cited in Kankaanpaa87: special designs for keyboard input for Kanji/Chinese

• [Murase85] (*p) Murase, H., Wakahara, T., and Umeda, M. "Online Writing-Box Free Character String Recognition by Candidate Character Lattice Method", Transactions of the Institute of Electronics Communications Engineers of Japan, Vol J68-D, pp 765-772, April 1985. (in Japanese).

  • Segmenting strokes into characters using varying spacing: box-less input

• [MurphyAS84] () Murphy, A.S. and Wiseman, J.A. "Optical Digitizing Stylus", IBM Technical Disclosure Bulletin, Volume 27, Number 5 October 1984, Pages 2806-2807

  • Cited in BennettWE91
  • Optical syslus digitizer using table with X and Y lines in different colors: appears to work like relative-position optical mouse, not absolute position

• [Murtagh84] (*a) Murtagh, F. and Raftery, A.E. "Fitting Straight Lines to Point Patterns", Pattern Recognition, Vol 17 NO 5, May 1984, pp 479-483.

  • Feature: finding straight lines in OCR patterns (skeletons?)

• [Nagayama85] (*p) Nagayama, T., Shibuya, J., and Kawakita, T. "Pen-Touch-Type Electro-Magnetic Transparent Touch Panel", Society for Information Display 1985 Digest of Technical Papers, May 1985,pp 32-35.

  • Transparent digitizer table over LCD with DCR

• [Nagy84] (*p) Nagy, G. and Seth, S. "Hierarchical Image Representation with Application to Optically Scanned Documents", Proceedings of the 7th International Conference on Pattern Recognition, Montreal, 1984.

  • Tiling/windows applied for format of written/printed documents

• [Nagy85] (*p) Nagy, G. "Image Database", Image and Vision Computing, Vol 3 No 3, August 1985, pp 111-117.

  • User-interface aspects of image data bases not given much attention

• [Naito84] (*a) Naito, S. and Masuda, T. "Chinese Character Recognition Based on Personal Handwriting Characteristics", Transactions of IECE of Japan, Vol J67D No 4, April 1984, pp 480-487 (Abstract only).

  • Science Citation Index
  • Difference in handwriting variations/variability between writers 50% greater than for one writer (?)
  • Claims 96.9% successful recognition when trained to writer, 67.7% for general use without training, for 3256 Kanji categories.

• [Nandhakumar85] (*p) Nandhakumar, N. and Aggarwal, J.K. "The Artificial Intelligence Approach to Pattern Recognition -- A Perspective and an Overview", Pattern Recognition, Vol 18 No 6, pp 383-389, 1985.

  • Overview of DCR in the context of artificial intelligence

• [NationalOA84] (*p) National OA: "Panaword RL-W450 / FW-500 product literature", (in Japanese), 1984.

  • National OA 1984 Kanji/Chinese product

• [NEC84] () NEC: "CR-100 product literature", (in Japanese), 1984.

  • Nippon Electric Company 1984 Japanese/Kanji/Chinese on-line handwriting product

• [Nouh84] (*p) Nouh, Adnan, Sultan, Abobakr, and Tolba, Roshdi "On Feature Extraction and Selection for Arabic Character Recognition", Arab Gulf Journal of Scientific Research, Vol 2 No 1, 1984, pp 329-347.

  • Density features for Arabic handwriting recognition: position-by-position matching between characters: Freeman codes as divergence criteria
  • Tilt: 0.01" to 20 degrees, 0.08" between 20 and 40 degrees

• [Numonics85] () Numonics "Numonics #2200 Pad with Low Profile and High Resolution", Numonics Incorporated, 418 Pierce Street, Lansdale PA 19446, 1985.

• [OedR84] (*p) Oed, R. and Doster, W. "The Intelligent Graphics Tablet - More Than a Mouse", Proceedings of Advances in Microprocessor Peripherals, Journees d'Electronique et de Microtechnique, Lausanne, Switzerland, October 8-11, 1984, pp 183-190.

  • Shows handwriting user-interface with control/editing area much like Penpad or C.I.C. Handwriter
  • Shows diagram for stroke/character segmentation/parsing using overlapping rectangular extents
  • User interface for handwriting into existing applications using keyboard buffer

• [OedR85] (*) Oed, R. and Doster, W. "On-line Script Recognition - A User-friendly Man-Machine Interface", Proceedings of the COMPINT 85 Conference on Computer-Aided Technologies, IEEE Catalog No 85CH2136-0, pp 741-743.

  • Shows diagram for stroke/character segmentation/parsing using surrounding rectangles. Separate writing areas on digitizer tablet for text input, gestures, virtual tablet/mouse input: compare with virtual tablets and recognition macros in Pencept PenPad. Makes statement that user-definable trainable/adaptable gesture-symbol recognition simplifies user interface, but this is questionable: semantics of symbol must still be defined, and mnemonicity and non-ambiguity (to recognizer) of a gesture is an open question.

• [OKI84] (*p) OKI Corporation "SR 220 / SR 230 Online product information", OKI Corporation, 1984.

  • Japanese handwriting character recognition product

• [OlsenDR84] (*p) Olsen, Dan. R. Jr., Buxton, William, Ehrich, Roger, Kasik, David J., Rhyne, James R., and Siberg, John "A Context for User Interface Management", IEEE Computer Graphics and Applications, December 1984, pp 33..41

• [Omron84] (*p) Omron Corporation "Handwriting product information", Omron Corporation, 1984.

  • Japanese handwriting character recognition product

• [Panati84] (*) Panati, Charles "Panati's Browser's Book Of Beginnings", Houghton Mifflin, Publishers, 1984. ISBN 0-395-56238-4

  • Chapter 6: "Scribble to Shorthand". Invention of Pens and Papyrus, historical reference to Stylus. Tieronian unistroke alphabet in the history of shorthand, 63 B.C., pages 81..82. Pitman Shorthand 1937, Taylor shorthand, Gregg Shorthand. Braille.

• [Parks85] (*) "Apparatus for Signature Verification", Parks, John R; Carr, David R; Fox, Peter F. United States Patent 4,495,644, January 22, 1985, assigned to Quest Automation Public Limited Company, Great Britain.

  • Quest Automation Handwriting product: Micropad/Datapad using resistive film, pressure-sensitive digitizer tablet with hand rest
  • Signature verification considering multiple scores of deviations, signature rejected either by a high combined score, or a very high score on one parameter

• [Parnell84] () Parnell "(title)", United States Patent 4,473,717

  • Lukis87 on digitizer tablets

• [Pavlidis85] () Pavlidis, J., and Van Wyk, C.J. "An Automatic Beautifier for Drawings and Illustrations", A.C.M. Computer Graphics, Vol 19 No 3, July 1985, pp 225-234.

  • Gesture/graphics input heuristics: cite for Pavlidis
  • Another spread-sheet for the Penpad

• [Pencept84a] () Pencept "Penpack (TM) Multiplan (TM) User's Guide", Pencept, Incorporated, 39 Green Street, Waltham, Massachusetts 02154, 1984.

  • Another word processor for the Penpad

• [Pencept84b] () Pencept "Penpack (TM) IBM (R) Personal Editor User's Guide", Pencept, Incorporated, 39 Green Street, Waltham, Massachusetts 02154, 1984.

  • Forms data entry for the Penpad

• [Pencept84c] () Pencept "Penware (TM) Penform (TM) User's Guide", Pencept, Incorporated, 39 Green Street, Waltham, Massachusetts 02154, 1984.

  • Pencept character set

• [Pencept84d] () Pencept "Users Manual for Penpad 320", Pencept, Incorporated, 39 Green Street, Waltham, Massachusetts 02154, 1984.

• [Pencept84e] () Pencept "Software Tool Kit for the Penpad 320", Pencept, Incorporated, 39 Green Street, Waltham, Massachusetts 02154, 1984.

• [Pencept84f] () Pencept "Penpad (R) 320 Technical Data", Pencept, Incorporated, 39 Green Street, Waltham, Massachusetts 02154, 1984.

  • Pencad, gesture / recognition macros for AutoCAD and other MS-DOS CAD applications.

• [Pencept84g] (*) Pencept "IBM PC peripheral accepts hand-printed or drawn input", IEEE CG&A, May 1984, "New Products", page 94.

  • Product announcement for Pencept PenPad 320, hand-writing-recognition terminal for IBM-PCs. Essentially a keyboard replacement with handwriting recognition, combined with a general purpose digitizing tablet. "Application Packs" for recognition macros / gestures using simulated keyboard/mouse input for unmodified Multiplan, Visical, Word Start, Lotus 1-2-3 spreadsheet. Included a simulated "mouse" mode, using either mouse input through the driver, or sending keyboard arrow-key input.

• [Pencept85a] () Pencept "Pencad (TM) Penpad (R) Interface for Freelance (R)", Pencept, Incorporated, 39 Green Street, Waltham, Massachusetts 02154, 1985.

• [Pencept85b] () Pencept "Pencad Penpad Interface for CAD and Graphics, Version 3.4 (R)", Pencept, Incorporated, 39 Green Street, Waltham, Massachusetts 02154, 1985.

  • shows recognition macros / gestures

• [PepperW84] () Pepper, William Jr. "Hand-held musical instrument and systems including a man-machine interface apparatus", United States Patent 4,430,917, February 14, 1984

• [Penycook85] (*) Pennycook, Bruce. W. "Computer-Music Interfaces: a Survey", Computing Surveys, Vol 17, No 2, June 1985, pp 267-ff

  • Mentions Groove and NRC music systems using light-pen, three-D wand digitizer
  • musical score-editing systems mentions in bibliography: some use pen/tablet input

• [Photron85] () Photron Limited "OSCON Digitizer technical information: SQ-3100 and SQ-4000MKII", Photron Limited, 1985.

• [Plamondon85] (*p) Plamondon, R. and Lamarche, F. "Modelization of Handwriting: A System Approach", in "Graphonomics: Contemporary Research in Handwriting", Proceedings of the 1986 Graphonomics Conference, Hong Kong, July 8-12, 1985, Henry S.R. Kao, Gerard P. van Galen and Rumjahn Hoosain, editors, North-Holland, publishers, 1986.

  • Analog filter on digitizer output gives "improved accuracy" (!!??!!)
  • Tablet digitizer jitter noise is +- 0.002 inch (?)

• [Polhemus84] () Polhemus "3-Space Digitizer Technical Bulletin", 1984, Polhemus Navigation Sciences Division of McDonnel-Douglass Electronics Co., Colchester, Vermont.

• [Poole84] () Poole, L. "A Tour of the Mac Desktop", MacWorld, Vol 1 No 1, 1984, pp 16-21.

  • Cited in JohnsonJ89

• [PortableComputer85] () Portable Computer Products Company "Data Pad product announcement", April 1985. 10801 West Main Street, Suite 600, Fairfax VA 22030, 703-352-8116 (see also Annographics). Mark Spikell, founder.

• [Prentice85a] () Prentice, Thomas C. "Methods for testing digitizer performance", Internal Report, Pencept, Incorporated, 1985.

• [Prentice85b] (*) Prentice, Thomas C. "Understanding Digitizer Specifications", Internal Report, Pencept, Incorporated, 1985.

• [Raeder85] () Raeder, G. "A Survey of Current Graphical Programming Techniques", IEEE Computer, August 1985, pp 11-25.

• [Richards85] (*p) Richards, Joh, T, Boies, Stephen J., Gould, John D. "Rapid Prototyping and System Development: Examination of an Interface Toolkit for Voice and Telephony Applications", IBM Research Division Report RC 11433 (#51395), 10/8/85, 9 pages

• [RoadsC85] (*) Roads, Curtis "Research in Music and Artificial Intelligence", Computing Surveys, Vol 17, No 2, June 1985, pp 163-ff

  • Editor of book including Buxton music-editing work

• [Rosson84] (*p) Rosson, M.B. "Characterizing Freeform Editing Behavior", IBM Research Report RC 10550 (Log #47200) 5/23/84, IBM T.J. Watson Research Center, PO Box 218, Yorktown Heights, NY 10598.

  • Boies said 75% of interactive user time is commands, not input: she says 25%
  • Common mistake is mis-typing command names
  • "direct manipulation" and gesture-input

• [Rouse85] (*p) Rouse, Nancy E. "Manual Digitizers for CAD", Machine Design, June 20, 1985

  • lists all major digitizer tablet companies of the day, and their technologies (sonic, magetic, magneto-strictive, etc.) 2-D: Altek, CalComp, Elographics, CTCO, Hitachi, Houston Instrument, Kurta, Numonics, Science Accessories Corp (SAC), Scriptel, Summagraphhics; 3-D: Micro Control Systems, Polhemus Navigation Sciences, Science Accessories Corp

• [Rubin85] (*p) Rubin, R.V., Golin, E.J., and Reiss, S.P. "ThinkPad: a Graphical System for Programming by Demonstration", IEEE Software, Vol 2 No 2, pp 73-79.

  • Interactive graphical interface language for programming
  • Acoustic 3-D tablet

• [SAC84] () SAC "GP-8-3D Grafbar Technical Bulletin", Science Accessories Corporation, 1984, Southport, Connecticut.

  • Acoustic 2-D digitizer tablet, large active area, refers to "standard digitizer output formats"

• [SAC88] (*p) SAC "GP-10 Two-dimensional Sonic Digitizer", Science Accessories Corporation, 1985(?), Southport, Connecticut.

  • Acoustic 2-D digitizer tablet, large active area, refers to "standard digitizer output formats": Emulation of vendors Kura, Numonics, Summagraphics, Wacom, Hitachi, Houston Instruments, GTCO, Calcomp and SAC GP-7, GP-8, CP-9, GP-9XL. Formats and behavior customizable via the controller, built-in menu in controller

• [Sakai84] (*p) Sakai, T., Odaka, K., and Toida, T. "Several Approaches to Development of On-line Handwritten Character Input Equipment", Proceedings of the 7th International Conference on Pattern Recognition, 1984, pp 1052-1054.

  • Boxed input: command-mode characters straddle boxes
  • New text-editing method for keyboard emulation: uses multiple lines
  • "category 1" characters: complete ambiguous due to sampling resolution

• [Sakoe84] (*p) Sakoe, Hiroaki "Pattern Matching Device Operable with Signals of a Compressed Dynamic Range", United States Patent 4,479,236, October 23, 1984, assigned to Nippon Electric Company, Tokyo, Japan.

  • Iterative approximation to distance function between two patterns

• [Satoh85] (*p) Satoh, K., Ooka, A., and Wada, Y. "Character Recognition Apparatus", United States Patent 4,542,526, September 17, 1985, assigned to Sumitomo Electric Industries, Osaka, Japan.

  • Handwriting character recognition using length and slope of line segments, classification to standard strokes
  • Uses 8-direction chain codes for directions of entire stroke segments. Uses distance between strokes to segment strokes into characters

• [Schaeken84] () Schaeken, B., Verschueren, W., Rene de Cotret, Y. and Hermanne, A. "A Hierarchical System for Handwritten Numeral Recognition", Proceedings of the 7th International Conference on Pattern Recognition, 1984, Montreal Canada, pp 1004-1007.

  • Says recognition success % available, but does not give it (!)
  • Handwriting recognition product

• [Schiermann84] (*) Schuermann, Juergen and Doster, Wolfgang "Umschreibende Rechtecke separieren Zeichen", Computerwoche Nr 46, 9 Nov 1984

  • Article on the various technologies used to move to a "paperless" office: optical scanning to convert paper documents to electronic form, handwriting/digitizers to capture character writing and sketsches without paper

• [Scriptel84] (*) Scriptel "New Products: CAD Graphics Tablet", IEEE Communications Magazine, Vol 22 No 4, April 1984, pp 50..51

  • Scriptel electrostatic tablet product advertisement: photograph shows transparent front-surface digitizer, suitable for digitizing drawing under the tablet, or frosted surface for rear-projection. Not subject to magnetic biasing (apparent reference to Summagraphics)

• [ScriptWriter85] (*p) Data Entry Systems "ScriptWriter Technical Information", Data Entry Systems, 6767 Madison Pike, Suite 195, Huntsville, Alabama 35806, 1985.

• [Shakunaga84] () Shakunaga, T, Kaneko, H., and Yodogawa, E. "Automatic Writer Recognition by Line-based Spectral Resolution of 2nd Order Statistics", Transactions of IECE of Japan, Vol J67D No 7, July 1984, pp 776-783 (in Japanese: English abstract only).

  • Science Citation Index
  • Like signature verification, but regular text handwriting: identification vs. verification, with 5.2% reject, 0.0034% false acceptance

• [Sharp84] (*p) Sharp Corporation "Handwriting product information", Sharp Corporation, 1984.

  • Japanese handwriting character recognition product

• [Shneiderman85] () Shneiderman, B. "Human-Computer Interaction Research at the University of Maryland", unpublished manuscript, October, 1985. Author's address: Department of Computer Science, Human-Computer Interaction Laboratory, University of Maryland, College Park, Maryland 20742.

  • User-interface studies on menu structure: globally visible menu helps
  • User-interface: arrow keys for pointing vs mouse
  • User-interface: manuals are better than on-line help (why?)
  • User-interface: user preference for touch-screen pointing tablet/display

• [Shridhar84a] () Shridhar, M. and Badreldin, A. "A tree classification algorithm for handwritten character recognition", Proceedings of the 7th International Joint Conference on Pattern Recognition, 1984, pp 615-618.

  • Cited in Kerrick88
  • Kerrick88 cites as reference on sequential, heirarchical, and tree-based methods for character recognition

• [Shridhar84b] () Shridhar, M. and Badreldin, A. "High accuracy character recognition algorithm using Fourier and topological descriptors", Pattern Recognition, Vol 17, 1984, pp 515-524.

  • 98% accuracy on handwriting recognition of numerics: Fourier on outlines confuses 2 & 5, 6 & 9, sometimes 4 & 7 due to rotational invariance, so topological analysis must follow.

• [Sklarew85] (*) Sklarew, Ralph "Linus Technologies Inc. Historical notes", Historical notes at www.blinkenlights.com

  • Write-Top introduced by Linus Technologies in 1987, mentions earlier work (ca 1984) with Robert Nadeau at George Mason University

• [Swan85] (*) Swan, K. "Pragmatics of Programming Voice Recognition Technologies", Proceedings of Voice I/O Systems Conference, San Francisco California, 1985.

  • Cited in Mangione86
  • Adaptive recognition: "trying to retraining an already overtrained behavior, such as speech, is not recommended"
  • Training/adaptive: users are unlikely to become expert users if they become discouraged naive users
  • Speech: human factors gap: most system designers and users weak in human factors/user interface issues
  • Speech: system must be readily interfaceable to external host applications
  • Speech: user do not want any "control knobs" on their user interface
  • Speech: features for good interface design never all on one product
  • Speech: user interface system designer cannot differentiate between user and machine error
  • Speech: products ignore transducer (microphone) problems, even though critical
  • Speech: writing custom software for applications kills the product
  • Speech: software must be transportable between programmers, not computers
  • Speech: buffering problems cause system to lose second word even though product works fine
  • Speech: speech equivalent of "graphical doodling"

• [Sziklai84] () Sziklai, G.C. "Methods and Apparatus for the Automatic Classification of Patterns", United States Patent 4,454,610, Jue 12, 1984, assigned to Transaction Sciences Corporation, New Orleans, Louisiana.

  • How to find the signature and courtesy amount on a check using a grey-scale image
  • Signature verification via OCR scanning

• [Tappert84a] () Tappert, C.C. "Dehooking Procedure for Handwriting on a Tablet", IBM Technical Disclosure Bulletin, Vol 27 No 5, October 1984, pp 2995-2998.

  • Hook removal: hooks are a problem of the tablet

• [Tappert84b] () Tappert, C.C. "Delayed Stroke Processor for Handwriting Recognition", IBM Technical Disclosure Bulletin, Vol 26 No 12, May 1984, pp 6616-6619.

  • Delayed strokes: aligns dots only with cusps, crosses on down side of cusp
  • Refers to character cut points in cursive character segmentation

• [Tappert84] (*) Tappert, C.C. "Adaptive On-line Handwriting Recognition", Proceedings of the 7th International Conference on Pattern Recognition, 1984, Montreal Canada, pp 1004-1007.

  • Refers to context by "alphabet selection"
  • Mentions U/V ambiguity for single writer (see also Jans75)
  • Mentions A/H, h/n, a/u ambiguity as problem of dynamic matching algorithm (!)
  • Ambiguity of 1,l,I, and upper/lower C,K,O,P,S,V,W,X,Y,Z
  • Mentions "new variations" even for single user (!)
  • Mentions sloppy and ambiguous writing as separate cases
  • Mentions "alphabet selection" like our shift boxes
  • Discrete vs continuous writing, but testing was different (biased)
  • Mentions problems with his hook removal algorithm
  • Claims 94% recognition success, 98% if restricted subset, with three familiar writers
  • Says dictionary lookup would give 100% recognition (!)
  • Mentions segmentation for case where text is known (but when is that real?)
  • Testing found several omitted variants from a "good" system

• [Tappert85a] () Tappert, C.C. "An Adaptive System for Handwriting Recognition", IBM Research Report RC 11175 (Log #50249) 5/21/85, 1985, IBM T.J. Watson Research Center, PO Box 218, Yorktown Heights, NY 10598 (to appear in Acta Psychologica).

  • Implies need for big data base: "vast variation, need larger quantities"
  • Refers to intra-writer variability
  • Claims at most four strokes in an English (Latin) letter
  • Mentions characters (C,c, 1,I,J, etc.) only distinguishable by context
  • Mentions need to accommodate new variations: "small sample of data"
  • Baseline drift, more writing time means more prototypes

• [Tappert85b] () Tappert, C.C. et al "Handwriting Recognition on Transparent Tablet over Flat Display", IBM Research Report RC 11856 (Log #52695) 3/3/86, 1986, IBM T.J. Watson Research Center, PO Box 218, Yorktown Heights, NY 10598. (To appear in 1986 Society for Information Display International Symposium Digest of Technical Papers).

  • Essential to success of character recognition is a natural human interface
  • Character recognition has biggest performance demands on digitizer tablets
  • Refers to commercially available handwriting recognizers, but not by name
  • Mentions some performance problems of commercially available tablets
  • Refers to minimum tablet tick resolution for character recognition
  • Mentions many tablet performance problems, "natural" writing speed

• [Tarnopolsky84] () Tarnopolsky "(title)", United States Patent 3,793,469

  • Lukis87 on digitizer tablets

• [ThomasJC85] () Thomas, John C. "Human Factors in IBM", IBM Thomas J. Watson Research Center, Report RC 11267 (#50753), July 15, 1985.

• [Toida83] . Toida, T. et al "A Method for Correcting Characters in On-Line Handwritten Input", Showa 58-nendo Denshi Tsushin Sogo Zenkoku Taikai, April 1983

  • Cited in Sklarew99 patent

• [VanDamm84] () Van Damm, A. "Scientific American", Introduction to special issue on computer technology, September 1984.

  • Apple MACINTOSH user-interface is really neat

• [VanRaamsdonk84] (*) Van Raamsdonk, Cornelis "Apparatus for Text Editing and Processing", United States Patent 4,475,239, October 2, 1984, assigned to Olympia Werke AG, Wilhelmshaven, West Germany.

  • Patent on user-interface for doing word processing with handwriting recognition, handwriting annotation input: prior art to just about every ink-editing system you can imagine. Editing of electronic ink.

• [Verschueren84] () Verschueren, W., Schaeken, B., Rene de Cotret, Y., and Hermanne, A. "Structure Recognition of Handwritten Numerals", Proceedings of IEEE 7th International Conference on Pattern Recognition, Montreal, Canada, July 30-August 2, 1984, pp 760-762.

  • Cites 6-0 ambiguity for single writer
  • Says 79% accuracy, "then improved", but does not say how much
  • Kerrick88 cites as example system using both locally and globally computed features

• [Wacom85a] (*) WACOM Corporation "Equipment/Device: Wacom CU-510 Power Supply for Wacom digitizer tablet, ca. 1985", WACOM Corporation

  • Physical device in collection.

• [Wakahara84a] () Wakahara, T. and Umeda, M. "On-line Cursive Script Recognition using Stroke Linkage Rules", Proceedings of IEEE 7th International Conference on Pattern Recognition, Montreal, Canada, July 30-August 2, 1984, pp 1065-1068.

• [Wakahara84b] () Wakahara, T., Odaka, K., and Umeda, M. "Electro-Luminescent Display Superposed by Transparent Electromagnetic Coupling Tablet and Its Application to Script Input Japanese Word Processor", (in Japanese), 1984.

  • Integrated DCR tablet and display, electronic ink
  • Display parallax corrected by user-selectable offset in electronic ink

• [Waltz85] () Waltz, D.L. and Pollack, J.B. "Massively parallel parsing", Cognitive Science, Vol 9, 1985, pp 51-74.

  • Feldman88

• [WangPSP85] (*) Wang, P.S.P. "Dual Level Pattern Recognition System", United States Patent 4,521,909, June 4, 1985, assigned to Wang Laboratories, Inc., Lowell, Massachusetts.

  • Two-level character recognition: a first cut may or may not work, but a second pass sorts out characters left ambiguous by the first pass
  • Template matching with learning for new patterns
  • Two-level template matching using a coarse digitizing grid, then a fine grid for what's left: only advantage is shorter comparison time

• [Ward85] () Ward, J. and Blesser, B. "Interactive Recognition of Hand-printed Characters for Computer Input", IEEE Computer Graphics and Applications, September 1985, pp 24-37.

• [Ward85a] . Ward, Jean Renard "Software Control at the Stroke of a Pen", SIGGRAPH Video Review, ACM Press: CHI '86.

  • Videotape of live demonstration during CHI '85 conference showing gesture recognition / recognition macros, handwriting recognition, interfaces to AutoCAD CAD and other applications.

• [Ward85b] (*p) Ward, J. and Blesser, B. "Implications of Using Interactive Hand-print Character Recognition for Computer Input", Proceedings of the 1985 COMPINT Trends and Applications in Computer Graphics Conference, May 1985, IEEE Catalog No 85CH2148-5.

  • Pre-cursor to Ward85 IEEE article: only cite Ward85, not this one

• [Ward85c] () Ward, J. "UNIX as a development tool for a non-UNIX microprocessor", CommUNIXations, August/September 1985, Vol V No 5, pp 26-30.

• [Ward85d] () Ward, J. "Method and Apparatus for Removing Noise at the Ends of a Stroke", United States Patent 4,534,060, August 6, 1985, assigned to Pencept, Incorporated, Waltham, Massachusetts.

  • Eliminate wobbles at ends of strokes for pre-processing handwritten characters

• [Ward85e] (*) Ward, J. and Nilssen, A. "Apparatus and Method for Emulating Computer Keyboard Input with a Hand-print Terminal", United States Patent 4,562,304, assigned to Pencept, Incorporated, Waltham, Massachusetts, December 31, 1985.

  • Keyboard emulation patent intended to blockanybody from doing on-line character recognition with standard keyboard input

• [Ward85f] (*p) Ward, Jean Renard "Pencept's Challenge", /usr/group CommUNIXactions, August/September 1985

  • engineering article on hardware/firmware development effort for the PenPad 300 product

• [Watanabe85] (*p) Watanabe, Y., Gyoba, J., Hirata, T., and Maruyama, K. "A Psychological Approach to the Human Recognition of Ambiguous Characters", Journal of the Institute of TV Engineers of Japan, Vol 39 No 6, pp 509-515, 1985 (in Japanese).

  • Mentions different "correct" labels for readers and writers of characters
  • Doing same stuff as early Shillman/Blesser work, but for Japanese
  • Refers to "confoundable Katakana" characters
  • "Hayashi's Quantification, model II"

• [Weibel84] (*) Weibel, Peter "Zur Geschichte und Aesthetik der digitalen Kunst", Supplement to catalog of Ars Electronica 84, Linz Austria, September 1984

  • History of computer graphics as artistic medium (mostly animation), first section a discussion of the technical implementation. States that for graphical artistic input, a data rate of 24 points/second on a digitizing tablet with a stylus is usually enough.

• [Werth85a] (*p) Werth, L.J. and Paulson, L.G. "Pattern Processing System", United States Patent 4,541,115, September 10, 1985, assigned to Pattern Processing Technologies, Minneapolis, Minnesota.

  • Visual image inspection pattern matching using address loop correlation between images: recognition part of a trainable system
  • Visual pattern image recognition: responds to time-varying images (patterns over time: claims application to speech recognition as well)

• [Werth85] (*p) Werth, L.J. and Paulson, L.G. "Training Controller for Pattern Processing System", United States Patent 4,504,970, March 12, 1985, assigned to Pattern Processing Technologies, Minneapolis, Minnesota.

  • Visual image inspection pattern matching using address loop correlation between images: trainable system

• [Weyer85] () Weyer, S.A. and Borning, A.H. "A prototype electronic encyclopedia", ACM Transactions on Office Information Systems", Vol 3 No 1, 1985, pp 63-88.

  • BarkerPG89: See also ThingLab

• [Whetton85] () Whetton, Cris, quoted in "A.C.M. Software Engineering Notes", Vol 5 No 10, pp 9-10, regarding an incident where published curves for a control system were transcribed from photocopies, October, 1985.

  • Xerox-copy error, for tablets, optical scanners

• [Wong85] (*p) Wong, K.H. and Fallsize, F. "Dynamic Programming in the Recognition of Connected Handwritten Script", Proceedings of IEEE Conference on Artificial Intelligence Applications, 1985, pp 666-670.

  • Tests limited to only one writer
  • Script recognition: complains about segmentation mismatches
  • Context: says storing templates for all possible words not practical

• [WuWT84] () Wu, W.T. "Heuristic Approach to Chinese-Character Search", Kybernetics (Great Britain), Vol 13 No 1, 1984, pp 39-42 (abstract only).

• [Yankelovich85] () Yankelovich, H., Meyrowitz, N., and van Dam, A. "Reading and writing the electronic book", IEEE Computer, Vol 18 No 10, pp 15-30, 1985.

  • Barker,PG89

• [Ye84] (*p) Ye, P.J., Hugli, H., and Pellandini, F. "Techniques for On-line Chinese Character Recognition with Reduced Writing Constraints", Proceedings of the 7th International Conference on Pattern Recognition, Montreal, 1984, pp 1043-1045.

  • On Chinese, 99% accuracy if use stroke order/sequence, 92% if do not
  • Test protocol: says extrapolation from 500 to 2500 Chinese characters straightforward

• [YouZ84] () You, Z. and Jain, A.K. "Performance evaluation of shape matching via chord length distribution", Computer Vision and Graphical Information Processing, Vol 29, 1984, pp 185-198.

  • Taxt90

• [Zhang84] (*p) Zhang, T.Y and Suen, C.Y. "A Fast Parallel Algorithm for Thinning Digital Patterns", Communications of the A.C.M., Vol 27 No 3, March 1984, pp 236-239.

  • Skeletonization/thinning based on opposing corners and boundaries along diagonals, preserving endpoints and pixel connectivity

• [Zue85] (*p) Zue, V.W. "The Use of Speech Knowledge in Automatic Speech Recognition", Proceedings of IEEE, Vol 73 No 11, November 1985, pp 1602-1614.

  • General-purpose pattern recognition not extensible to read world of multiple speakers, large vocabulary, continuous speech
  • Phonetically-based approaches abandoned for adaptive recognition because no one could deal with phoneme detection well
  • General-purpose pattern recognition appeals because it is simple, but it does not scale up to real problems
  • Variability in speech due to co-articulation of phonemes
  • Human reading of spectrograms: speech segmentationbetter on nonsense words than on regular text (!)
  • Knowledge of allophonic variationshelpful for phonetic decoding
  • Use known properties of human auditory system to reduce irrelevant acoustic variabilities
  • Much recognition to be learned from studying human decoding of speech signal
  • Synthesis (artificial generative variability) methods no help on analysis/recognition
  • Stressed syllables are "islands of reliability" in speech segmentation
  • Recognition dichotomy of template matching vs feature extraction not absolute
  • Template matching works well on small, distinct set, but not extensible to real world
  • Speech signal is a special kind of signal
  • Ignorance: correctly defining what you do not know is critical: tells you what to work on to make the best progress
  • Ignorance: admit what you do not know, then do something about it

      Main Reference Page