Showing papers on "Handwriting recognition published in 2000"

Online and off-line handwriting recognition: a comprehensive survey

Abstract: Handwriting has continued to persist as a means of communication and recording information in day-to-day life even with the introduction of new technologies. Given its ubiquity in human transactions, machine recognition of handwriting has practical significance, as in reading handwritten notes in a PDA, in postal addresses on envelopes, in amounts in bank checks, in handwritten fields in forms, etc. This overview describes the nature of handwritten language, how it is transduced into electronic data, and the basic concepts behind written language recognition algorithms. Both the online case (which pertains to the availability of trajectory data during writing) and the off-line case (which pertains to scanned images) are considered. Algorithms for preprocessing, character and word recognition, and performance with practical systems are indicated. Other fields of application, like signature verification, writer authentification, handwriting learning tools are also considered.

Method and system for providing alternatives for text derived from stochastic input sources

Abstract: A computer-implemented method for providing a candidate list of alternatives for a text selection containing text from multiple input sources, each of which can be stochastic (such as a speech recognition unit, handwriting recognition unit, or input method editor) or non-stochastic (such as a keyboard and mouse). A text component of the text selection may be the result of data processed through a series of stochastic input sources, such as speech input that is converted to text by a speech recognition unit before being used as input into an input method editor. To determine alternatives for the text selection, a stochastic input combiner parses the text selection into text components from different input sources. For each stochastic text component, the combiner retrieves a stochastic model containing alternatives for the text component. If the stochastic text component is the result of a series of stochastic input sources, the combiner derives a stochastic model that accurately reflects the probabilities of the results of the entire series. The combiner creates a list of alternatives for the text selection by combining the stochastic models retrieved. The combiner may revise the list of alternatives by applying natural language principles to the text selection as a whole. The list of alternatives for the text selection is then presented to the user. If the user chooses one of the alternatives, then the word processor replaces the text selection with the chosen candidate.

Biometric personal identification based on handwriting

Abstract: In this paper, we describe a new method to identify the writer of Chinese handwritten documents. There are many methods for signature verification or writer identification, but most of them require segmentation or connected component analysis. They are content dependent identification methods, as signature verification requires the writer to write the same text (e.g. his name). In our new method, we take the handwriting as an image containing some special texture, and writer identification is regarded as texture identification. This is a content independent method. We apply the well-established 2D Gabor filtering technique to extract features of such textures and a weighted Euclidean distance classifier to fulfil the identification task. Experiments are made using Chinese handwritings from 17 different people and very promising results were achieved.

Method and apparatus for multiple simultaneously active data entry mechanisms on a computer system

Abstract: A method and system providing simultaneous data entry for a computer system having both on-screen keyboard entry and mechanisms for handwriting recognition entry. In one embodiment, a portable or palmtop computer system contains a flat panel display screen capable of displaying thereon a keyboard image (“virtual keyboard”). Characters can be entered into the computer system by a user interacting with (e.g., tapping) the displayed characters of the virtual keyboard. The computer system also provides a handwriting recognition mechanism (e.g., digitizer pad) whereby characters are recognized based on a user drawing strokes on the pad. In accordance with the present invention, the virtual keyboard and the handwriting recognition mechanism are simultaneously active for data entry. Therefore, the computer system can accept character entry from the handwriting recognition mechanism while the virtual keyboard is displayed and active and capable of providing character entry itself. Specifically, from the virtual keyboard, the user is allowed to either tap the individual buttons, representing characters, on the screen to enter data or the user can enter data via the handwriting recognition mechanism. Since both methods are active simultaneously, the user does not have to switch between them. The present invention gives a user increased flexibility in the manual entry of characters to the computer system.

A Line-Oriented Approach to Word Spotting in Handwritten Documents

Abstract: The problem of word spotting in handwritten archives is approached by matching global shape features. A set of visual templates is used to define the keyword class of interest, and initiate a search for words exhibiting high shape similarity to the model set. Major problems of segmenting cursive script into individual words are avoided by applying line-oriented processing to the document pages. The use of profile-oriented features facilitates the application of dynamic programming techniques to pattern matching, and allows us to achieve high levels of recognition performance. Results of experiments with old Spanish manuscripts show a high recognition rate of the proposed approach.

Handwriting recognition method and apparatus having multiple selectable dictionaries

Abstract: Multiple selectable dictionaries are used by a handwriting recognition system to provide accurate and prompt recognition processing. Sets of the multiple dictionaries are selectable by the user or automatically by the recognition program in response to the user handwriting in predetermined fields on a user interface.

Generalized hidden Markov models. II. Application to handwritten word recognition

Abstract: For part I see ibid. vol.8, no. 1 (2000). This paper presents an application of the generalized hidden Markov models to handwritten word recognition. The system represents a word image as an ordered list of observation vectors by encoding features computed from each column in the given word image. Word models are formed by concatenating the state chains of the constituent character hidden Markov models. The novel work presented includes the preprocessing, feature extraction, and the application of the generalized hidden Markov models to handwritten word recognition. Methods for training the classical and generalized (fuzzy) models are described. Experiments were performed on a standard data set of handwritten word images obtained from the US Post Office mail stream, which contains real-word samples of different styles and qualities.

Computer system with integrated telephony, handwriting and speech recognition functions

Abstract: A computer system with speech recognition system (24) and handwriting recognition system (25) is disclosed that work together to improve the total recognition accuracy of each alone. The handwriting recognition system may include a pen/stylus input device (10) and associated program function. The system may be combined with computer and telephone functions (8) to provide an integrated application having voice output programs (26), Internet access (22), e-mail/v-mail (29) and personal information manager (23) functions. The system can recognize speaker dependent and speaker independent speech, converting this information to computer recognizable text, which may be displayed onto a display device (78) in near real-time. Speech recognition errors may be corrected via a pen input device (10), and the pen information may be recognized, converted to text and graphics. This data may then be displayed at near real-time or displayed later at a user specified time. The recognized handwritten pen information may be integrated into the speech recognized text and stored in a data storage system (18).

Multimodal meeting tracker

Abstract: Face-to-face meetings usually encompass several modalities including speech, gesture, handwriting, and person identification. Recognition and integration of each of these modalities is important to create an accurate record of a meeting. However, each of these modalities presents recognition difficulties. Speech recognition must be speaker and domain independent, have low word error rates, and be close to real time to be useful. Gesture and handwriting recognition must be writer independent and support a wide variety of writing styles. Person identification has difficulty with segmentation in a crowded room. Furthermore, in order to produce the record automatically, we have to solve the assignment problem (who is saying what), which involves people identification and speech recognition. We follow a multimodal approach for people identification to increase the robustness (with the modules: color appearance id, face id and speaker id). This paper will examine a meeting room system under development at Carnegie Mellon University that enables us to track, capture and integrate the important aspects of a meeting from people identification to meeting transcription. Once a multimedia meeting record is created, it can be archived for later retrieval. This paper will review our meeting browser that we have developed which facilitates tracking and reviewing meetings.

Digital pen using speckle tracking

Abstract: A digital pen that has an ink writing tip includes a laser on a pen body that directs light toward paper across which the writing tip is stroked. A CMOS camera or CCD is also mounted on the pen body for detecting reflections of the laser light, referred to as “speckles”. A processor in the pen body determines relative pen motion based on the speckles. A contact sensor such as an FSR on the pen body senses when the tip is pressed against the paper, with positions being recorded on a flash memory in the pen body when the contact sensor indicates that the pen is against the paper. The memory can be later engaged with a handwriting recognition device to correlate the positions to alpha-numeric characters. Ordinary paper can be used, but, if desired, special bar-coded paper can also be used, so that the recorded positions can be tagged with a page number, form field, and absolute position on the page.

Integrated handwriting and speed recognition systems

Abstract: A computer system with speech recognition system and handwriting recognition system are disclosed that work closely together to improve the total recognition accuracy of each alone. The handwriting recognition system may include a pen/stylus input device and associated program functions. The system or programs may be combined with computer telephony functions to provide intergrated applications having voice output programs, Internet access, e-mail/v-mail and personal information manager functions. The computer system can recognize speaker-dependent and speaker-independent speech, converting this information to computer recognizable text, which may be displayed onto a display device in near realtime. Speech recognition errors may be corrected via a pen input device, and the pen information may be recognized, converted to text and graphics. This data may then be displayed at near realtime or displayed later at a user specified time. Recognized handwritten pen information may be intergrated into the speech recognized text and stored in a data storage system.

Method and apparatus for backlighting a handwriting input area for a portable computing device

Abstract: A method and apparatus for backlighting a handwriting input area for a portable computing device. The portable computing device includes a display area for displaying alphanumeric data and other images. Underneath the display area is a digitizer input area by which users enter handwritten information into the portable computing device. The portable computing device contains handwriting recognition software which converts the handwritten information into alphanumeric data. Both the display area and the digitizer input area are both backlit to facilitate usage in poor lighting conditions.

Graphical handwriting recognition user interface

Abstract: A graphical handwriting recognition user interface, or GUI (10), keyboard button (26), includes a display (LCD) and processor (102), one or more areas designated on said display for enabling entry of handwritten information using a stylus, and an image of a character being displayed within the one or more areas designated for entering the handwritten information, the image depicting a form of the handwritten information to be entered. The image of character is displayed in a manner depicting the character being entered, thereby rendering animation to the image of the character being displayed.

Towards a multimodal meeting record

Abstract: Face-to-face meetings usually encompass several modalities including speech, gesture, handwriting, and person identification. Recognition and integration of each of these modalities is important to create an accurate record of a meeting. However, each of these modalities presents recognition difficulties. Speech recognition must be speaker and domain independent, have low word error rates, and be close to real time to be useful. Gesture and handwriting recognition must be writer independent and support a wide variety of writing styles. Person identification has difficulty with segmentation in a crowded room. Furthermore, in order to produce the record automatically, we have to solve the assignment problem (who is saying what), which involves people identification and speech recognition. This paper examines a multimodal meeting room system under development at Carnegie Mellon University that enables us to track, capture and integrate the important aspects of a meeting from people identification to meeting transcription. Once a multimedia meeting record is created, it can be archived for later retrieval.

Online handwriting recognition using multiple pattern class models

Abstract: The field of personal computing has begun to make a transition from the desk-top to handheld devices, thereby requiring input paradigms that are more suited for single hand entry than a keyboard and recent developments in online handwriting recognition allow for such input modalities. Data entry using a pen forms a natural, convenient interface. The large number of writing styles and the variability between them makes the problem of writer-independent unconstrained handwriting recognition a very challenging pattern recognition problem. The state-of-the-art in online handwriting recognition is such that it has found practical success in very constrained problems. In this thesis, a method of identifying different writing styles, referred to as lexemes, is described. Approaches for constructing both non-parametric and parametric classifiers are described that take advantage of the identified lexemes to form a more compact representation of the data, while maintaining good recognition accuracies. Experimental results are presented on different sets of unconstrained online handwritten characters and words. In addition, a method of combining information from lexeme models built on different feature sets is described, and results are presented on both English characters and Devanagari characters. Finally, a method of writer-adaptation is described which makes use of the lexemes identified from a large group of writers to define lexemes within a small amount of data from a single writer.

Digital pen using visible image and autocorrelation of substrate pattern

Abstract: A digital pen that has an ink writing tip includes a light source on a pen body that directs light toward paper across which the writing tip is stroked. A CMOS camera or CCD is also mounted on the pen body for detecting reflections of the light. A processor in the pen body determines relative pen motion based on the reflections. A contact sensor such as an FSR on the pen body senses when the tip is pressed against the paper, with positions being recorded on a non-volatile memory in the pen body only when the contact sensor indicates that the pen is against the paper. Periodically, key frames are stored in memory but not for every cycle. The memory can be later engaged with a handwriting recognition device, such as a PC, to correlate the key frames and positions to alpha-numeric characters. Ordinary paper or quad-ruled paper can be used, and, if desired, special bar-coded paper can also be used, so that the PC can determine absolute pen position.

A discrete hmm for online handwriting recognition

Abstract: A fast HMM algorithm is proposed for online handwritten character recognition. After preprocessing the input strokes are discretized so that a discrete HMM is used. This particular discretization naturally leads to a simple procedure for assigning the initial state and state transition probabilities. In the training phase, complete marginelization with respect to state is not performed. A criterion based on normalized maximum likelihood ratio is given for deciding when to create a new model for the same character in the learning phase, in order to cope with stroke order variations and large shape variations. Experiments are done on the Kuchibue data base from Tokyo University of Agriculture and Technology. The algorithm appears to be very robust against stroke number variations and have reasonable robustness against stroke order variations and large shape variations. A drawback of the proposed algorithm is its memory requirement when the number of character classes and their associated models becomes large. Density tying is discussed in order to overcome this problem.

Online signature verification using pen-position, pen-pressure and pen-inclination trajectories

Abstract: A new algorithm is proposed for pen-input online signature verification incorporating pen-position, pen-pressure and pen-inclinations trajectories. The preliminary experimental result obtained looks encouraging.

Pen computing: challenges and applications

Abstract: Pen computing as a fieId broadly includes computers and applications in which a pen is the main input device. This field continues to draw a lot of attention from researchers because there are a number of applications where the pen is the most convenient form of input. These include: 1. preparing a first draft of a document and concentrating on content creation; 2. a socially acceptable form of capturing information in meetings, that is quieter than typing and creates minimal visual barrier; 3. applications that need privacy; 4. entering letters in ideographic languages like Chinese and Japanese and non-letter entries like graphics, music and gestures; and 5. interaction with multi-modal systems. The advent of electronic tablets in the late 1950s precipitated considerable activity in the area of pen computing. This interest ebbed in the 1970's, and was renewed in the 1980's, primarily due to advances in pen hardware technology and improvement in user-interfaces and handwriting recognition algorithms. There are still however, a number of challenges that need to be addressed before pen computing can address the needs listed above to a acceptable level of user satisfaction. In the paper, an overview of three aspects of pen computing are presented: pen input hardware, handwriting recognition and pen computer applications.

Support vector machine and its application in handwritten numeral recognition

Abstract: The support vector machine (SVM) is a new learning machine with very good generalization ability, which has been applied widely in pattern recognition and regression estimation. Four practical handwritten numeral SVM classifiers are proposed in this paper, which has been utilized successfully in Chinese check recognition system. The experiment on NIST numeral database and the actual check samples show that compared with other classifiers, SVM possesses better generalization ability.

Handwriting recognition-the last frontiers

Abstract: The last frontiers of handwriting recognition are considered to have started in the last decade of the second millennium. The paper summarizes (a) the nature of the problem of handwriting recognition, (b) the state of the art of handwriting recognition at the turn of the new millennium, and (c) the results of CENPARMI researchers in automatic recognition of handwritten digits, touching numerals, cursive scripts, and dates formed by a mixture of the former 3 categories. Wherever possible, comparable results have been tabulated according to techniques used, databases, and performance. Aspects related to human generation and perception of handwriting are discussed. The extraction and usage of human knowledge, and their incorporation into handwriting recognition systems are presented. Challenges, aims, trends, efforts and possible rewards, and suggestions for future investigations are also included.

Handwritten address interpretation: A task of many pattern recognition problems

Abstract: A gradation of pattern discrimination problems is encountered in interpreting handwritten postal addresses. There are several multiclass discrimination problems, including handwritten numeral recognition with 10 classes, alphanumeral recognition with 36 classes, and touching-digit pair recognition with 100 classes. Word recognition with a lexicon is a problem where the number of classes varies from a few to about a thousand. Some of the discrimination techniques, particularly those with few classes, lend themselves well to neural network classification, while others are better handled by Bayesian polynomial and nearest-neighbor methods. This paper describes each of the discrimination problems and the performances of each of the subsystems in a handwritten address interpretation system developed at CEDAR.

Online writing-box-free recognition of handwritten Japanese text considering character size variations

Abstract: An online writing-box free method for recognizing handwritten Japanese text is proposed. This method is achieved by the following procedure. First, the average character size of input handwritten text is estimated. Second, candidates for character segmentation are detected using geometric features between two adjacent strokes. Finally, a search is performed by dynamic programming for the string that maximizes evaluation score (acceptability as Japanese text). The evaluation score reflects the likelihood of character segmentation, recognition, context and the size of each character. The size is a newly introduced factor since alphabets, numerals, symbols, Japanese phonetic characters, simple Chinese characters and compound Chinese characters are all written in different sizes even in a single line of text. Experimental results show that the incorporation of the character size likelihood increases the character recognition rate for Japanese text.

User adaptive handwriting recognition by self-growing probabilistic decision-based neural networks

Abstract: Based on self-growing probabilistic decision-based neural networks (SPDNNs), user adaptation of the parameters of SPDNN is formulated as incremental reinforced and anti-reinforced learning procedures, which are easily integrated into the batched training procedures of the SPDNN. In this study, we developed: 1) an SPDNN based handwriting recognition system; 2) a two-stage recognition structure; and 3) a three-phase training methodology for a global coarse classifier (stage 1), a user independent hand written character recognizer (stage 2), and a user adaptation module on a personal computer. With training and testing on a 600-word commonly used Chinese character set, the recognition results indicate that the user adaptation module significantly improved the recognition accuracy. The average recognition rate increased from 44.2% to 82.4% in five adapting cycles, and the performance could finally increase up to 90.2% in ten adapting cycles.

A model of stroke extraction from Chinese character images

Abstract: Given the large number and complexity of Chinese characters, pattern matching based on structural decomposition and analysis is believed to be necessary and essential to off-line character recognition. The paper proposes a model of stroke extraction for Chinese characters. One problem for stroke extraction is how to extract primary strokes. Another major problem is to solve the segmentation ambiguities at intersection points. We use the degree information and the stroke continuation property to tackle these two problems. The proposed model can be used to extract strokes from both printed and handwritten character images.

Automatic handwriting gestures recognition using hidden Markov models

Abstract: Hidden Markov models have been successfully employed in speech recognition and, more recently, in sign language interpretation. They seem adequate for visual recognition of gestures. In this paper, two problems often eluded are considered. We propose to use the Bayesian information criterion in order to determine the optimal number of model states. We describe the contribution of continuous models in opposition to symbolic ones. Experiments on handwriting gestures show recognition rate between 88% and 100%.

Off-line unconstrained handwritten word recognition

Abstract: In this paper, we describe our system for writer independent, off-line unconstrained handwritten word recognition. We have developed a new method to automatically determine the parameters of Gabor filters to extract features from slant and tilt corrected images. An algorithm is also developed to translate 2D images to 1D domain. Finally, we propose a modified dynamic programming method with fuzzy theory to recognize words. Our initial experiments have shown promising results.