WO2000057349A1 - Handwriting recognition system - Google Patents
Handwriting recognition system Download PDFInfo
- Publication number
- WO2000057349A1 WO2000057349A1 PCT/GB2000/001052 GB0001052W WO0057349A1 WO 2000057349 A1 WO2000057349 A1 WO 2000057349A1 GB 0001052 W GB0001052 W GB 0001052W WO 0057349 A1 WO0057349 A1 WO 0057349A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- signals
- sampling
- classifier
- handwriting recognition
- recognition system
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/14—Image acquisition
- G06V30/142—Image acquisition using hand-held instruments; Constructional details of the instruments
- G06V30/1423—Image acquisition using hand-held instruments; Constructional details of the instruments the instrument generating sequences of position coordinates corresponding to handwriting
Definitions
- the present invention relates to a handwriting recognition system and more particularly to a hardware and an algorithm for implementing such a system.
- PCT application number GB98/031 6 Publication No. W099/223378
- a portable computer in the form of a pen-type casing.
- Incorporated within the casing is at least one accelerometer which is used to detect movement of the pen with respect to its environment.
- a handwriting recognition system comprising means responsive to input analogue signals representative of movement of a handheld writing device, sampling means to provide signals representative of the acceleration of the writing device in at least x axis and y axis channels at a predetermined capture sampling rate, and filtering means to remove dc level components and to provide smoothing of the output whereby signals representative of movement of the pen over a period are supplied to a classifier for comparison with a template representative of characters formed.
- the classifier may use hidden Markov modelling (HMM) techniques using a large number of states to determine the character defined by movement.
- HMM hidden Markov modelling
- the system may include an input indicative of a user's intention that the movement is representative of character writing.
- a method of analysing signals from a moving handheld device comprising sampling signals at a predetermined rate, passing signals through a bandpass filter to remove dc level and excess acceleration components, sampling the filtered output to provide a series of vectors representing the position of the handheld device at periodic intervals and using a classifier to compare the sample sets with predetermined templates to determine the character for output.
- Figure 1 is a block schematic diagram of the system
- Figure 2 shows relative positioning of the x and y axis of the handwriting device of Figure 1 ;
- Figure 3 is a schematic diagram of the handheld writing device of Figure 1 in a particular position
- Figure 4 shows relative input and output vector streams from the system of Figure 1 ;
- Figures 5 to 9 show comparative templates for a number of different letters.
- an input device 1 such as a stylus produces x and y vector streams 2 and 3 which are fed into a sampling unit 4.
- the outputs x and y are generally from accelerometers or other position sensing devices within the stylus 1 .
- Also feeding the sampling unit 4 is an output 5 from a switch indicated here as being in the nib section of the stylus 1 such that contact between the switch 6 and a surface is indicative of the stylus being used in a writing mode.
- a switch indicated here as being in the nib section of the stylus 1 such that contact between the switch 6 and a surface is indicative of the stylus being used in a writing mode.
- the nib switch 6 when incorporated in a non-surface contacting stylus, such as that disclosed in the previously referred PCT application, may be replaced by a user operable switch.
- the x and y vectors are passed to a classifier 9, which uses a hidden Markov model to carry out a comparison between the vectors and templates representing written characters.
- the classifier 9 may be arranged to output to a visual display 10.
- the stylus 1 for example comprises a simple plastic casing containing the electronics for transferring information to a PC.
- Two accelerometers mounted in the stylus, for example, are used to produced the x and y outputs.
- the nib switch is a simple on/off switch connected to determine when pressure is being applied to the pen nib and can therefore detect when a pen for stylus 1 is writing.
- the two accelerometers mounted in the top of the pen measure acceleration across their plane such that effectively they measure acceleration along the x axis 1 1 and the y axis 1 2 of the writing surface 14.
- the acceleration measured by the sensors is made up of two components, acceleration due to gravity and the acceleration as a result of stylus movement. It will be appreciated that the acceleration due to gravity is always present, such that when the pen is exactly horizontal both sensors would measure acceleration of ⁇ g.
- the accelerometers are subject to Sin ⁇ x .g where ⁇ is the inclination angle of the stylus 1 .
- acceleration as a result of the pen moving is produced by the acceleration and deceleration effect as the user writes.
- the acceleration of the two sensors x tota ⁇ and y tota can be expressed as
- the two acceleration signals are read into a normal PC using an RS232 port and the binary switch signal by means of the games port of a sound card.
- the sampling section must sample sufficiently regularly to capture the movement of the pen but should not over-sample, which would result in a waste of processing and storage within the PC. It has been found satisfactory for the purposes of the current invention to sample at a rate of 60 Hz.
- the acceleration signals for each channel are read in as two byte words giving a dynamic range for each acceleration signal from 0 to 65535.
- the pen nib switch is similarly sampled at the same rate.
- the accelerometer signals are partly dependent on a component of the earth's gravitation field passing through the accelerometer of the stylus 1 . This results in an almost constant dc level present on the output corresponding to the average pen angle ⁇ while writing.
- the bandpass filter 7 is thus arranged to filter the signals from the two accelerometers to remove the offsets.
- the bandpass filter smoothes the output from the sensors thus correcting for instability introduced by the user so that the smoothed output from the sensors increases robustness and facilitates matching between the x and y vectors and stored templates.
- the pen nib switch or manually operable switch indicates whether the stylus 1 was being used in a writing mode or not.
- the down sampling process 8 uses the information to down sample acceleration samples and to retain only those when it was known that the stylus 1 was writing.
- the x axis and y axis vectors are as shown at 1 5 and 1 6, then the down sampling vectors need only be taken into account when the nib switch signal indicator 1 7 is high.
- the output from the down sampling process is a time series of two dimensional vectors x and y as indicated at 1 8 and 1 9.
- the vector stream is passed to the classifier stage 9 which takes in a series of vectors representing the acceleration measurements made within a given word. These are then compared to a set of templates which cover the range of words within the system vocabulary and the word which matches most closely with the unknown input word is deemed the recognised word.
- the classifier is a hidden Markov model. Such models have been widely used in speech recognition and using a large number of states in the hidden Markov model will give the best performance for corresponding handwriting recognition.
- the display 10 which displays the output from the PC allows display of a word, for example, on a screen.
- a typical single accelerometer output can be seen respectively for the letters c, b, f and h in Figures 5 to 8.
- the template developed here shows three entries on a single accelerometer for each of the letters.
- f, b, h and c are shown in comparison so that a suitable template for comparison may be derived. It will be appreciated that the combination of an x accelerometer trace and a y accelerometer trace will serve further to emphasise the difference between each input letter.
- Bandpass filtering in digital form to remove dc components and high frequency components increases the reliability of the recognition process and therefore the reliability of the interpretation of the stylus input 1 . It will be appreciated that where the stylus 1 carries other components, for example an internal processing arrangement, some of the functions may be transferred from the PC to the stylus 1 . All of the components of sampling, bandpass filtering, down sampling and classifying can be implemented in a suitable computer program.
Abstract
Description
Claims
Priority Applications (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
AU33098/00A AU3309800A (en) | 1999-03-24 | 2000-03-21 | Handwriting recognition system |
US09/914,262 US7054510B1 (en) | 1999-03-24 | 2000-03-21 | Handwriting recognition system |
EP00911105A EP1181665A1 (en) | 1999-03-24 | 2000-03-21 | Handwriting recognition system |
US11/250,378 US20060088215A1 (en) | 1999-03-24 | 2005-10-17 | Handwriting recognition system |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
EP99302270.6 | 1999-03-24 | ||
EP99302270 | 1999-03-24 |
Related Child Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US11/250,378 Division US20060088215A1 (en) | 1999-03-24 | 2005-10-17 | Handwriting recognition system |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2000057349A1 true WO2000057349A1 (en) | 2000-09-28 |
Family
ID=8241290
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/GB2000/001052 WO2000057349A1 (en) | 1999-03-24 | 2000-03-21 | Handwriting recognition system |
Country Status (4)
Country | Link |
---|---|
US (2) | US7054510B1 (en) |
EP (1) | EP1181665A1 (en) |
AU (1) | AU3309800A (en) |
WO (1) | WO2000057349A1 (en) |
Families Citing this family (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6549935B1 (en) | 1999-05-25 | 2003-04-15 | Silverbrook Research Pty Ltd | Method of distributing documents having common components to a plurality of destinations |
US6816274B1 (en) * | 1999-05-25 | 2004-11-09 | Silverbrook Research Pty Ltd | Method and system for composition and delivery of electronic mail |
US20080174550A1 (en) * | 2005-02-24 | 2008-07-24 | Kari Laurila | Motion-Input Device For a Computing Terminal and Method of its Operation |
US7983478B2 (en) | 2007-08-10 | 2011-07-19 | Microsoft Corporation | Hidden markov model based handwriting/calligraphy generation |
TWI374391B (en) * | 2008-05-27 | 2012-10-11 | Ind Tech Res Inst | Method for recognizing writing motion and trajectory and apparatus for writing and recognizing system |
US8683582B2 (en) | 2008-06-16 | 2014-03-25 | Qualcomm Incorporated | Method and system for graphical passcode security |
TWI457793B (en) * | 2008-08-08 | 2014-10-21 | Ind Tech Res Inst | Real-time motion recognition method and inertia sensing and trajectory |
US9501694B2 (en) | 2008-11-24 | 2016-11-22 | Qualcomm Incorporated | Pictorial methods for application selection and activation |
GB2496790B (en) * | 2010-07-06 | 2019-01-16 | T Data Systems S Pte Ltd | Data storage device with data input function |
FR2969779A1 (en) | 2010-12-23 | 2012-06-29 | Movea | SYSTEM FOR SEIZING GRAPHIC ELEMENTS |
WO2015059930A1 (en) * | 2013-10-25 | 2015-04-30 | Wacom Co., Ltd. | Dynamic handwriting verification, handwriting-baseduser authentication, handwriting data generation, and handwriting data preservation |
CN104063705B (en) * | 2014-06-05 | 2017-08-11 | 北京捷通华声语音技术有限公司 | The method and apparatus that a kind of handwriting characteristic is extracted |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4736445A (en) * | 1986-01-21 | 1988-04-05 | International Business Machines Corporation | Measure of distinguishability for signature verification |
WO1997016799A2 (en) * | 1995-10-31 | 1997-05-09 | Baron Technologies Ltd. | Continuous security system based on motion code |
WO1997044758A1 (en) * | 1996-05-23 | 1997-11-27 | Apple Computer, Inc. | Methods and apparatuses for handwriting recognition |
Family Cites Families (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US3818443A (en) * | 1972-04-28 | 1974-06-18 | Burroughs Corp | Signature verification by zero-crossing characterization |
JPH06266490A (en) * | 1993-03-12 | 1994-09-22 | Toshiba Corp | Information input device and position recognition system for information input |
US5577135A (en) * | 1994-03-01 | 1996-11-19 | Apple Computer, Inc. | Handwriting signal processing front-end for handwriting recognizers |
US5781661A (en) * | 1994-06-29 | 1998-07-14 | Nippon Telegraph And Telephone Corporation | Handwritting information detecting method and apparatus detachably holding writing tool |
US5854855A (en) * | 1994-09-09 | 1998-12-29 | Motorola, Inc. | Method and system using meta-classes and polynomial discriminant functions for handwriting recognition |
US5802205A (en) * | 1994-09-09 | 1998-09-01 | Motorola, Inc. | Method and system for lexical processing |
US5768417A (en) * | 1994-09-09 | 1998-06-16 | Motorola, Inc. | Method and system for velocity-based handwriting recognition |
-
2000
- 2000-03-21 WO PCT/GB2000/001052 patent/WO2000057349A1/en not_active Application Discontinuation
- 2000-03-21 EP EP00911105A patent/EP1181665A1/en not_active Withdrawn
- 2000-03-21 AU AU33098/00A patent/AU3309800A/en not_active Abandoned
- 2000-03-21 US US09/914,262 patent/US7054510B1/en not_active Expired - Fee Related
-
2005
- 2005-10-17 US US11/250,378 patent/US20060088215A1/en not_active Abandoned
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4736445A (en) * | 1986-01-21 | 1988-04-05 | International Business Machines Corporation | Measure of distinguishability for signature verification |
WO1997016799A2 (en) * | 1995-10-31 | 1997-05-09 | Baron Technologies Ltd. | Continuous security system based on motion code |
WO1997044758A1 (en) * | 1996-05-23 | 1997-11-27 | Apple Computer, Inc. | Methods and apparatuses for handwriting recognition |
Also Published As
Publication number | Publication date |
---|---|
AU3309800A (en) | 2000-10-09 |
EP1181665A1 (en) | 2002-02-27 |
US7054510B1 (en) | 2006-05-30 |
US20060088215A1 (en) | 2006-04-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20060088215A1 (en) | Handwriting recognition system | |
TW403877B (en) | Handwriting input apparatus using more than one sensing technique | |
US6906703B2 (en) | Electronic module for sensing pen motion | |
Amma et al. | Airwriting: a wearable handwriting recognition system | |
Yanay et al. | Air-writing recognition using smart-bands | |
Ardüser et al. | Recognizing text using motion data from a smartwatch | |
CN106716317A (en) | Method and apparatus for addressing touch discontinuities | |
JPH11338626A (en) | Handwriting inputting method, handwriting input system and writing material | |
WO1996039677A1 (en) | Method and apparatus for character recognition of hand-written input | |
US6625314B1 (en) | Electronic pen device and character recognition method employing the same | |
Oh et al. | Inertial sensor based recognition of 3-D character gestures with an ensemble classifiers | |
CA2022075C (en) | Cross-product filter | |
US20150116285A1 (en) | Method and apparatus for electronic capture of handwriting and drawing | |
Milner | Handwriting recognition using acceleration-based motion detection | |
US9195887B2 (en) | Retrieving apparatus, retrieving method, and computer program product | |
Ji et al. | A new effective wearable hand gesture recognition algorithm with 3-axis accelerometer | |
Wang et al. | Real-time continuous gesture recognition with wireless wearable imu sensors | |
CN115769177A (en) | System for detecting handwriting problems | |
EP0542557B1 (en) | Systems and methods for handprint recognition acceleration | |
Teja et al. | A ballistic stroke representation of online handwriting for recognition | |
JPH1153104A (en) | Pen-type input device and pattern recognition method for the pen-type input device | |
Tuncer et al. | Handwriting recognition by derivative dynamic time warping methodology via sensor-based gesture recognition. | |
JP3712835B2 (en) | Pen-type input device | |
JPH11296292A (en) | Pen type input device and method for recognizing character | |
JP2018067132A (en) | Handwritten character recognizing device, detection device, and processing device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AK | Designated states |
Kind code of ref document: A1 Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BY CA CH CN CR CU CZ DE DK DM DZ EE ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX NO NZ PL PT RO RU SD SE SG SI SK SL TJ TM TR TT TZ UA UG US UZ VN YU ZA ZW |
|
AL | Designated countries for regional patents |
Kind code of ref document: A1 Designated state(s): GH GM KE LS MW SD SL SZ TZ UG ZW AM AZ BY KG KZ MD RU TJ TM AT BE CH CY DE DK ES FI FR GB GR IE IT LU MC NL PT SE BF BJ CF CG CI CM GA GN GW ML MR NE SN TD TG |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
DFPE | Request for preliminary examination filed prior to expiration of 19th month from priority date (pct application filed before 20040101) | ||
WWE | Wipo information: entry into national phase |
Ref document number: 09914262 Country of ref document: US |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2000911105 Country of ref document: EP |
|
REG | Reference to national code |
Ref country code: DE Ref legal event code: 8642 |
|
WWP | Wipo information: published in national office |
Ref document number: 2000911105 Country of ref document: EP |
|
WWW | Wipo information: withdrawn in national office |
Ref document number: 2000911105 Country of ref document: EP |