US20120239397A1 - Digital Ink Database Searching Using Handwriting Feature Synthesis - Google Patents
Digital Ink Database Searching Using Handwriting Feature Synthesis Download PDFInfo
- Publication number
- US20120239397A1 US20120239397A1 US13/482,828 US201213482828A US2012239397A1 US 20120239397 A1 US20120239397 A1 US 20120239397A1 US 201213482828 A US201213482828 A US 201213482828A US 2012239397 A1 US2012239397 A1 US 2012239397A1
- Authority
- US
- United States
- Prior art keywords
- digital ink
- handwriting
- text
- feature vectors
- query
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Abandoned
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/3331—Query processing
- G06F16/3332—Query translation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
-
- 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
- G06V30/22—Character recognition characterised by the type of writing
- G06V30/226—Character recognition characterised by the type of writing of cursive writing
Definitions
- the present invention broadly relates to digital processor implemented handwriting searching or recognition systems, and in particular, to a method of and apparatus for searching of a digital ink database using handwriting feature synthesis from a search query in text form.
- Digital ink database refers to a database which stores handwritten characters, for example a string of handwritten characters forming a handwritten letter.
- Pen-based computing systems provide a convenient and flexible means of human-computer interaction. Most people are very familiar with using pen and paper. This familiarity is exploited by known systems which use a pen-like device as a data entry and recording mechanism for text, drawings or calculations which are quite naturally supported by this medium. Additionally, written ink is a more expressive format than digital text, and ink-based systems can be language-independent.
- Pen-based computing allows users to store data in the form of notes and annotations, and subsequently search this data based on hand-drawn queries.
- searching handwritten text is more difficult than traditional text (e.g. ASCII text) searching due to inconsistencies in the production of handwriting and the stylistic variations between writers.
- the traditional method of searching handwritten data in a digital ink database is to first convert the digital ink database and corresponding search query to standard text using pattern recognition techniques, and then to match the query text with the converted standard text in the database. Fuzzy text searching methods have been described, see P. Hall and G. Dowling, “Approximate String Matching”, Computing Surveys, 12(4), pp. 381-402, 1980, that perform text matching in the presence of character errors, similar to those produced by handwriting recognition systems.
- handwriting recognition accuracy remains low, and the number of errors introduced by handwriting recognition (both for the database entries and for the handwritten query) means that this technique does not work well.
- the process of converting handwritten information into text results in the loss of a significant amount of information regarding the general shape and dynamic properties of the handwriting. For example, some letters (e.g. ‘u’ and ‘v’, ‘v’ and ‘r’, ‘f’ and ‘t’, etc.) are handwritten with a great deal of similarity in shape. Additionally, in many handwriting styles (particularly cursive writing), the identification of individual characters is highly ambiguous.
- Vallabhaneni “The Handwritten Trie: Indexing Electronic Ink”, The 1995 ACM SIGMOD International Conference on Management of Data , San Jose, Calif., May 1995; W, Aref, D. Barbera, D. Lopresti, and A. Tomkins, “Ink as a First-Class Datatype in Multimedia Databases”, Database System—Issues and Research Direction , pp. 113-163, 1996; and, R. Manmatha, C. Han, E. Riseman, and W. Croft, “Indexing Handwriting Using Word Matching”, Proceedings of the First ACM International Conference on Digital Libraries , pp. 151-159, 1996.
- the digital ink database searching techniques previously described are dependent on an ink query that is generated by the writer who authored the digital ink database. However, it would be beneficial if a digital ink database was searchable using other input mechanisms, for example, using a text query entered with a computer keyboard, or spoken and recognized by a voice recognition system. Alternatively, a third party may wish to search the digital ink database, either using his or her own handwriting, or using a text-based query.
- Ink database searching using handwriting feature synthesis allows a digital ink database to be searched using a text-based query.
- a text query is converted into feature vectors that are similar to the feature vectors that would have been extracted had the author of the digital ink database written the text query by hand.
- the feature vectors are then used to search the database, for example by using traditional techniques. This allows the searching of a digital ink database when the only input mechanism available is text entry, and can allow a person other than the author of the digital ink database to search the digital ink database.
- a method of searching a digital ink database comprising the steps of:
- FIG. 1 illustrates a processing system
- FIG. 2 illustrates an overview of the handwriting recognition method.
- FIG. 3 illustrates the digital ink database searching method using feature synthesis.
- FIG. 4 illustrates a third party ink database searching method.
- FIG. 5 illustrates training and recognition/search phases.
- FIG. 6 illustrates a method for handwriting recognition.
- FIG. 7 illustrates an example for text recognition.
- the present invention provides a method and apparatus for searching a digital ink database using a text query.
- like reference numerals are used to identify like parts throughout the figures.
- the processing system 10 generally includes at least a processor 20 , a memory 21 , and an input device 22 , such as a graphics tablet and/or keyboard, an output device 23 , such as a display, coupled together via a bus 24 as shown.
- An external interface is also provided as shown at 25 , for coupling the processing system to a digital ink database 11 .
- the processing system 10 is adapted to allow data to be stored in and/or retrieved from the digital ink database 11 .
- the processor 20 receives handwriting data, a text query, etc., via the input 22 . From this, it will be appreciated that the processing system 10 may be any form of processing system or terminal such as a computer, a laptop, server, specialised hardware, or the like.
- the writer-specific handwriting model describes the handwriting style of a particular user. Most user-adaptive handwriting recognition systems generate some kind of model to account for the stylistic variations between individual users. Generally, the purpose of these handwriting models is to map the feature vectors extracted from a set of input pen strokes into a set of letters that represent the recognized text.
- FIG. 2 details the general steps in a handwritten text recognition system.
- the handwriting is sampled 26 and raw ink 27 is passed to a normalization step 28 .
- the normalized ink 29 undergoes a segmentation step 30 and the resulting strokes 31 are passed to a feature extraction step 32 which extracts the feature vectors 33 .
- Classification step 34 is then performed using the handwriting model 35 , which produces the primitives 36 .
- Text recognition step 37 receives the primitives 36 and uses the language model 38 and/or handwriting model 35 to generate text 39 corresponding to raw ink 27 .
- the writer-specific handwriting model 35 is required to be modified to store information that allows text to be mapped to ink features.
- a table (called a lexicon) translating individual letters and groups of letters (i.e. sequences of letters) into features is included in the handwriting model. After recognition has been performed, letters in the output text and the corresponding feature vectors used for recognition can be added to the lexicon.
- Groups of commonly co-occurring letters and their corresponding feature vectors are concatenated and added to the lexicon. This is desirable since handwriting exhibits co-articulation effects (where the writing of a letter is influenced by the shape of surrounding letters) and commonly written letters (e.g. “qu”, “ed”, and “ing”) are likely to exhibit co-articulation.
- a more accurate rendering of the query ink can be produced that considers the contextual effects of the stroke sequence.
- the lexicon should be able to store multiple feature vectors for each letter sequence. Since the same letters can be recognized many times, the lexicon should be able to select the feature vectors that best represents the letters. This can be done by storing all feature vector sequences output by the recognizer for each letter sequence, together with a count of the number of times that feature vector sequence has been seen for that letter sequence. The feature vector sequence with the highest count (i.e. the most frequently seen and thus most probable) for a letter sequence is then used during feature synthesis.
- letter groups can be stored based on character transition statistics (for example, as derived from a text corpus) where character sequences with a high probability of occurrence are stored (e.g. the probability of “ing” is much greater than the probability of “inx”).
- character transition statistics for example, as derived from a text corpus
- all possible letter groupings can be stored after recognition, with some kind of culling procedure executed when the table becomes too large (for example, least-recently used).
- the text query is converted to a set of feature vectors by the feature synthesis procedure that uses the writer-specific handwriting model. These feature vectors are then used as the query term for searching the digital ink database.
- the ink database search can be performed using traditional ink matching techniques.
- FIG. 3 describes this procedure.
- text is input and the text 41 provided to feature synthesis step 42 which uses the handwriting model 35 to generate features 43 .
- the features 43 are used in an ink searching step 44 of the digital ink database 11 . This produces ink matches 45 .
- a search of the lexicon is performed to locate a sequence of lexicon entries that contain letter sequences that can be combined to produce the query text.
- the stored feature vectors are concatenated to produce the query feature vector.
- a modified A* search algorithm see S. Russell and P. Norvig, Artificial Intelligence—A Modern Approach , Prentice Hall, 1995, can be used to search for a letter sequence s, where the path cost function g(s) is the number of lexicon entries used so far to create the text, and the estimated cost to the goal is:
- This heuristic states that at least one additional lexicon entry is required to complete the letter sequence if the sequence has fewer letters than the query word. Nodes in the search tree are sorted by g(s)+h(s) (with lower scores being superior) and nodes with the same score are ordered by the number of letters in the sequence (with higher being superior).
- Each row in the table represents a search node, with higher scoring nodes located at the top of the table:
- the top node is now a completed sequence, and no other node in the search can produce a better score, so this node is selected as the search result.
- Ink searching using feature synthesis can be performed without using a handwriting recognition system.
- the technique only requires the ability to build a lexicon of letter sequences and associated feature vectors to model the handwriting of the writer who authored the digital ink database.
- a training procedure can be used to generate the writer-specific lexicon. To do this, a user provides a sample of their handwriting by copying a specified training text, which is then used to build the lexicon.
- the training procedure is not required to perform full handwriting recognition since the text represented by the handwriting is already known; rather, it simply needs to segment the input into characters and strokes, convert the strokes into features, and store the appropriate letter groups and associated feature vectors in the lexicon.
- the training text used to build the lexicon should be optimized to provide a balanced example set of individual letters and letter groups. That is, it should maximize the coverage of likely character unigrams, bigrams, and trigrams, see J. Pitrelli, J. Subrahmonia, M. Perrone, and K. Nathan, “Optimization of Training Texts for Writer-Dependent Handwriting Recognition”, Advances in Handwriting Recognition , World Scientific Publishing, 1999, with an emphasis on letters and letter sequences that are most likely to be encountered.
- both the handwriting recognition system and the ink-matching algorithm use the same feature representations, since the features used to search the digital ink database are derived from the results of the handwriting recognition.
- Some feature sets can allow a transformation from recognition features to search features to be learned from a set of training data.
- many feature sets can allow the regeneration of an approximation of the digital ink from the recognition features, from which the second feature set can be extracted. That is, the text query is converted to a set of feature vectors using feature synthesis, and the inverse transformation of the feature-extraction process is applied to the features to convert them into digital ink from which the search features are extracted. Care must be taken to ensure that this procedure does not introduce artifacts in the digital ink that may affect the search feature extraction (e.g. discontinuities in the generated ink may cause problems for some feature extraction techniques).
- a person can search the digital ink database of another writer by using handwriting recognition to convert their ink input into text, and using feature synthesis to convert the recognized text into features for ink searching.
- FIG. 4 depicts this situation, with writer B searching the digital ink database authored by writer A.
- ink is received from writer B and the strokes 47 are passed to feature extraction step 48 .
- the features 49 are extracted and recognition step 50 uses a language model 51 and writer B model 52 to generate corresponding text 53 . This is used to build a lexicon for writer B at lexicon builder step 54 . Letter groups and features are sent back to writer B model 52 to improve/update the model.
- the text 53 then undergoes feature synthesis step 42 and a similar process as that described with reference to FIG. 3 is then followed to retrieve ink matches 45 that have been authored by writer A.
- the procedure for handwriting recognition and ink searching using feature synthesis requires a training phase and a recognition or searching phase.
- a training phase a set of training data is converted to stroke features that are clustered into stroke primitives and used to create a model of the handwriting.
- this model is used to decode input ink or synthesize features for ink searching. This process is depicted in FIG. 5 . Note that the same preprocessing, normalization, segmentation, and feature extraction procedure is used for training, recognition and searching.
- the handwriting recognition system maps input ink into a set of stroke code words that are used to search a dictionary for matching words.
- FIG. 6 provides an overview of this process, with the individual steps described below in more detail.
- the ink is sampled at a constant rate of 100 Hz.
- Research has shown that handwriting has a peak spectral density at around 5 Hz that declines to noise level at about 10 Hz, see H. Teulings and F. Maarse, “Digital Recording and Processing of Handwriting Movements”, Human Movement Science, 3, pp. 193-217, 1984.
- a low-pass filter with a cutoff at 10 Hz will remove the high-frequency noise without affecting the relevant spectral components of the handwriting signal.
- a low-pass filter conforming to the above specifications can be produced by circularizing the point coordinates, performing an FFT to remove the high frequency components, and recreating the signal using an inverse-FFT.
- a simple weighted averaging filter works as effectively.
- the filter width k and ⁇ smoothing coefficients are determined empirically.
- a i angle ⁇ ( p i , p i - 1 )
- Angles ⁇ 1 and ⁇ 2 define which stroke segments are considered as parts of a down stroke and are empirically set at 40° and 140° respectively (with 90° representing a vertical line). If the estimated slant deviates more than a certain threshold from the vertical, slant is removed using a shear transformation:
- x i ( y m ⁇ ⁇ ax - y i ) ⁇ tan ⁇ ( 90 ⁇ ° - slant ) y ma ⁇ ⁇ x - y m ⁇ ⁇ i ⁇ ⁇ n
- y min and y max represents the top and bottom of the bounding rectangle of the ink.
- Zone estimation is used to normalize the height of the input ink.
- English letters exhibit three zones—the middle zone (corresponding to the height of letters such as ‘a’, ‘c’, ‘e’, etc.), and the upper and lower zones that contain the ascenders and descenders in letters such as ‘b’, ‘d’, ‘g’, and ‘j’.
- Zone estimation is performed using a horizontal histogram of ink density. That is, the number of ink crossings is determined for an equally spaced series of horizontal lines passing through the bounding rectangle of the ink.
- the central peak of the histogram is found, as are the two points on either side of the histogram where the ink density drops below a certain fraction of the central peak height. These two points are selected as the upper and lower bounds of the middle zone.
- the upper and lower zones are defined as the space between the middle zone and the vertical extremum of the bounding rectangle.
- the ink is segmented into sub-strokes at extrema in the vertical direction (i.e. at local maxima and minima of the Y coordinates).
- the lengths of the two sub-strokes produced by splitting the stroke at the selected segmentation point must exceed a pre-calculated minimum distance (set at half the height of the estimated middle zone).
- segmented sub-strokes are then re-sampled to contain a constant n number of points located equidistantly along the stroke trajectory.
- a feature vector is then created for the sub-stroke by normalizing the coordinates:
- x 1 ′ x i - x m ⁇ ⁇ i ⁇ ⁇ n h
- y i ′ y i - y middle h
- the resulting vectors describe the sub-strokes using a large number of highly correlated features (clearly, the coordinates of point p i are dependent on point p i-1 and so on).
- the Karhunen-Loeve transform is used. This procedure projects the higher-dimensionality features into a lower dimension using linear mapping that is optimal in a least-squares sense.
- the eigenvectors and eigenvalues for this matrix are found (using the tridiagonal QL implicit algorithm, see W. Press, B. Flannery, S. Teukolsky, and W. Vetterling, W. T., Numerical Recipes in C , Cambridge: Cambridge University Press, 1988), and the eigenvectors corresponding to the largest n eigenvalues are used to form the PCA matrix Z.
- the feature vectors are then multiplied by this matrix to transform the features into a new feature space with an orthogonal uncorrelated axis:
- the transformed feature vectors are then clustered using a Kohonen Self-Organizing Feature Map (SOFM), see T. Kohonen, “Self-Organized Formation of Topologically Correct Feature Maps”, Biological Cybernetics, 43, pp. 59-69, 1982.
- SOFM Kohonen Self-Organizing Feature Map
- This technique uses an unsupervised learning procedure to cluster the input vectors such that distance and proximity relationships between the vectors is maintained as far as possible.
- the SOFM used has a 2-dimensional structure, with visually similar code words (i.e. clusters) located near each other. As a result, the distance between two code words can be easily calculated using some distance measure (e.g. Euclidean distance) between the code word values.
- SOFM training is performed iteratively using a simple two-layer neural network that is initialized with random weights.
- the best matching output neuron for a normalized input training vector x is found using minimum Euclidean distance:
- o ⁇ ( x ) argmin n i ⁇ ⁇ x - w i ⁇
- w i represents the weight vector of output node i.
- the weights of the node with the highest activation value and those nodes surrounding it (as determined by a neighborhood function ⁇ ) are updated using:
- ⁇ is a learning rate function and ⁇ and ⁇ are typically varied over time t. Training continues until there is no noticeable change to the neuron weights during an iteration of the training set.
- each sub-stroke feature vector is quantized using the trained SOFM and appended to the code word vector.
- a feature vector is quantized into a code word by selecting the largest activation value for the output neurons of the SOFM codebook:
- the handwriting model stores a mapping of stroke code word vectors to letters (for text recognition) and the reverse mapping of letter groups to code word vectors (for feature synthesis). To build the handwriting model, each letter in the training data is converted to a code word vector that is stored in a table, along with the corresponding letter.
- a specific code word sequence may map to a number of letters (e.g. poorly-drawn ‘u’ may map to the same feature vectors as a ‘v’) and individual letters may be mapped to by a number of code word vectors.
- the probability that the vector represents the letter x i can be calculated given n lexicon entries for letter x i :
- This table indicates that if the code word sequence ⁇ 3,4 ⁇ is encountered in the input, there is a probability of 0.54 that it represents the letter ‘u’, a probability 0.41 that it represents an ‘v’, and a probability of 0.05 that it represents an ‘r’.
- the reverse mapping table is produced in a similar way, storing code word vectors associated with letters and groups of letters.
- the input ink is processed as described above, and the resulting code word vectors are used to search the handwriting model to produce letter hypotheses.
- a best-first search strategy is used to search a dictionary for the most likely word given the derived letter probabilities. The process is depicted below in FIG. 7 .
- Ink searching is performed by using the handwriting model generated during handwriting recognition to map the input query text into a sequence of code words.
- This code word vector is used to search the digital ink database using an elastic matching technique.
- a full description of a similar ink-matching technique is given in D. Lopresti and A. Tomkins, “Temporal-Domain Matching of Hand-Drawn Pictorial Queries”, Handwriting and Drawing Research: Basic and Applied Issues , IOS Press, pp. 387-401, 1996.
- the resulting queries are then ordered by similarity and presented to the user.
- the invention may also be said broadly to consist in the parts, elements and features referred to or indicated in the specification of the application, individually or collectively, in any or all combinations of two or more of said parts, elements or features, and where specific integers are mentioned herein which have known equivalents in the art to which the invention relates, such known equivalents are deemed to be incorporated herein as if individually set forth.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- Educational Administration (AREA)
- Tourism & Hospitality (AREA)
- Computational Linguistics (AREA)
- Multimedia (AREA)
- Game Theory and Decision Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Development Economics (AREA)
- General Business, Economics & Management (AREA)
- Character Discrimination (AREA)
- Image Analysis (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Transition And Organic Metals Composition Catalysts For Addition Polymerization (AREA)
- Burglar Alarm Systems (AREA)
- Regulating Braking Force (AREA)
- Position Input By Displaying (AREA)
- Analysing Materials By The Use Of Radiation (AREA)
Abstract
A method of searching a digital ink database is disclosed. The digital ink database is associated with a specific author. The method starts by receiving a computer text query from an input device. The computer text query is then mapped to a set of feature vectors using a handwriting model of that specific author. As a result, the set of feature vectors approximates features that would have been extracted had that specific author written the computer query text by hand. Finally, the set of feature vectors is used to search the digital ink database.
Description
- The present application is a Continuation application of U.S. application Ser. No. 10/492,169 filed on Apr. 9, 2004 which is a National Phase Application of PCT/AU02/0395, filed on Oct. 15, 2002, all of which is herein incorporated by reference.
- The present invention broadly relates to digital processor implemented handwriting searching or recognition systems, and in particular, to a method of and apparatus for searching of a digital ink database using handwriting feature synthesis from a search query in text form.
- “Digital ink database” as used herein refers to a database which stores handwritten characters, for example a string of handwritten characters forming a handwritten letter.
- Pen-based computing systems provide a convenient and flexible means of human-computer interaction. Most people are very familiar with using pen and paper. This familiarity is exploited by known systems which use a pen-like device as a data entry and recording mechanism for text, drawings or calculations which are quite naturally supported by this medium. Additionally, written ink is a more expressive format than digital text, and ink-based systems can be language-independent.
- The increasing use of pen computing and the emergence of paper-based interfaces to networked computing resources (for example see: P. Lapstun, Netpage System Overview, Silverbrook Research Pty Ltd, 6 Jun., 2000; and, Anoto, “Anoto, Ericsson, and Time Manager Take Pen and Paper into the Digital Age with the Anoto Technology”, Press Release, 6 Apr., 2000), has highlighted the need for techniques which are able to store, index, and search (raw) digital ink. Pen-based computing allows users to store data in the form of notes and annotations, and subsequently search this data based on hand-drawn queries. However, searching handwritten text is more difficult than traditional text (e.g. ASCII text) searching due to inconsistencies in the production of handwriting and the stylistic variations between writers.
- The traditional method of searching handwritten data in a digital ink database is to first convert the digital ink database and corresponding search query to standard text using pattern recognition techniques, and then to match the query text with the converted standard text in the database. Fuzzy text searching methods have been described, see P. Hall and G. Dowling, “Approximate String Matching”, Computing Surveys, 12(4), pp. 381-402, 1980, that perform text matching in the presence of character errors, similar to those produced by handwriting recognition systems.
- However, handwriting recognition accuracy remains low, and the number of errors introduced by handwriting recognition (both for the database entries and for the handwritten query) means that this technique does not work well. The process of converting handwritten information into text results in the loss of a significant amount of information regarding the general shape and dynamic properties of the handwriting. For example, some letters (e.g. ‘u’ and ‘v’, ‘v’ and ‘r’, ‘f’ and ‘t’, etc.) are handwritten with a great deal of similarity in shape. Additionally, in many handwriting styles (particularly cursive writing), the identification of individual characters is highly ambiguous.
- Various techniques for directly searching and indexing a digital ink database are known in the prior art, see for example: A. Poon, K. Weber, and T. Cass, “Scribbler: A Tool for Searching Digital Ink”, Proceedings of the ACM Computer-Human Interaction, pp. 58-64, 1994; I. Kamel, “Fast Retrieval of Cursive Handwriting”, Proceedings of the 5th International Conference on Information and Knowledge Management, Rockville, Md. USA, Nov. 12-16, 1996; W. Aref, D. Barbera, P. Vallabhaneni, “The Handwritten Trie: Indexing Electronic Ink”, The 1995 ACM SIGMOD International Conference on Management of Data, San Jose, Calif., May 1995; W, Aref, D. Barbera, D. Lopresti, and A. Tomkins, “Ink as a First-Class Datatype in Multimedia Databases”, Database System—Issues and Research Direction, pp. 113-163, 1996; and, R. Manmatha, C. Han, E. Riseman, and W. Croft, “Indexing Handwriting Using Word Matching”, Proceedings of the First ACM International Conference on Digital Libraries, pp. 151-159, 1996.
- These systems use a similarity measure to compare a feature vector derived from a set of query pen strokes with a database of feature vectors derived from the digital ink database. The entries in the database that exhibit the greatest degree of similarity with the query are returned as matches. Additionally, some approaches create an index or use a partitioning scheme to avoid a sequential search of all entries in the database. See for example: D. Barbara, W. Aref, I. Kamel, and P. Vallabhaneni, “Method and Apparatus for Indexing a Plurality of Handwritten Objects”, U.S. Pat. No. 5,649,023; D. Barbara and I. Kamel, “Method and Apparatus for Similarity Matching of Handwritten Data Objects”, U.S. Pat. No. 5,710,916; D. Barbara and H. Korth, “Method and Apparatus for Storage and Retrieval of Handwritten Information”, U.S. Pat. No. 5,524,240; D. Barbara and W. Aref, “Method for Indexing and Searching Handwritten Documents in a Database”, U.S. Pat. No. 5,553,284; R. Hull, D. Reynolds, and D. Gupter, “Scribble Matching”, U.S. Pat. No. 6,018,591; A. Poon, K. Weber, and T. Cass, “Searching and Matching Unrecognized Handwriting”, U.S. Pat. No. 5,687,254; and, W. Aref and D. Barbara, “Trie Structure Based Method and Apparatus for Indexing and Searching Handwritten Databases with Dynamic Search Sequencing”, U.S. Pat. No. 5,768,423.
- Other studies, J. Hollerbach, “An Oscillation Theory of Handwriting”, Biological Cybernetics, pp. 139-156, 1981, and, Y. Singer and N. Tishby, “Dynamical Encoding of Cursive Handwriting”, IEEE Conference on Computer Vision and Pattern Recognition, 1993, describe efforts to model the physical properties of handwriting for handwriting synthesis.
- The digital ink database searching techniques previously described are dependent on an ink query that is generated by the writer who authored the digital ink database. However, it would be beneficial if a digital ink database was searchable using other input mechanisms, for example, using a text query entered with a computer keyboard, or spoken and recognized by a voice recognition system. Alternatively, a third party may wish to search the digital ink database, either using his or her own handwriting, or using a text-based query.
- Ink database searching using handwriting feature synthesis allows a digital ink database to be searched using a text-based query. Using a writer-specific handwriting model derived from a handwriting recognition system or suitable training procedure, a text query is converted into feature vectors that are similar to the feature vectors that would have been extracted had the author of the digital ink database written the text query by hand. The feature vectors are then used to search the database, for example by using traditional techniques. This allows the searching of a digital ink database when the only input mechanism available is text entry, and can allow a person other than the author of the digital ink database to search the digital ink database.
- According to an aspect of the present invention, there is provided a method of searching a digital ink database, the digital ink database being associated with a predetermined author, the method comprising the steps of:
-
- receiving a computer text query from an input device;
- mapping by a processing system the computer text query to a set of feature vectors using a handwriting model of the predetermined author, the set of feature vectors approximating features that would have been extracted had the predetermined author written the computer query text by hand; and
- using the set of feature vectors to search the digital ink database.
- Other aspects are also disclosed.
- The present invention should become apparent from the following description, which is given by way of example only, of a preferred but non-limiting embodiment thereof, described in connection with the accompanying figures, wherein:
-
FIG. 1 illustrates a processing system. -
FIG. 2 illustrates an overview of the handwriting recognition method. -
FIG. 3 illustrates the digital ink database searching method using feature synthesis. -
FIG. 4 illustrates a third party ink database searching method. -
FIG. 5 illustrates training and recognition/search phases. -
FIG. 6 illustrates a method for handwriting recognition. -
FIG. 7 illustrates an example for text recognition. - The following modes are described as applied to the written description and appended claims in order to provide a more precise understanding of the subject matter of the present invention.
- The present invention provides a method and apparatus for searching a digital ink database using a text query. In the figures, incorporated to illustrate the features of the present invention, like reference numerals are used to identify like parts throughout the figures.
- Embodiments of the present invention can be realised using a processing system an example of which is shown in
FIG. 1 . In particular, theprocessing system 10 generally includes at least aprocessor 20, amemory 21, and aninput device 22, such as a graphics tablet and/or keyboard, anoutput device 23, such as a display, coupled together via abus 24 as shown. An external interface is also provided as shown at 25, for coupling the processing system to adigital ink database 11. - In use, the
processing system 10 is adapted to allow data to be stored in and/or retrieved from thedigital ink database 11. Theprocessor 20 receives handwriting data, a text query, etc., via theinput 22. From this, it will be appreciated that theprocessing system 10 may be any form of processing system or terminal such as a computer, a laptop, server, specialised hardware, or the like. - The writer-specific handwriting model describes the handwriting style of a particular user. Most user-adaptive handwriting recognition systems generate some kind of model to account for the stylistic variations between individual users. Generally, the purpose of these handwriting models is to map the feature vectors extracted from a set of input pen strokes into a set of letters that represent the recognized text.
- However, the feature synthesis approach to digital ink searching uses a writer-specific handwriting model to perform the inverse mapping. That is, the model is used to convert query text into a set of feature vectors that approximate the features that would have been extracted had the writer written the query text by hand.
FIG. 2 details the general steps in a handwritten text recognition system. The handwriting is sampled 26 andraw ink 27 is passed to anormalization step 28. The normalizedink 29 undergoes asegmentation step 30 and the resultingstrokes 31 are passed to afeature extraction step 32 which extracts thefeature vectors 33.Classification step 34 is then performed using thehandwriting model 35, which produces theprimitives 36.Text recognition step 37 receives theprimitives 36 and uses thelanguage model 38 and/orhandwriting model 35 to generatetext 39 corresponding toraw ink 27. - To use feature synthesis for ink database searching, an additional step is required. The writer-
specific handwriting model 35 is required to be modified to store information that allows text to be mapped to ink features. To perform this mapping, a table (called a lexicon) translating individual letters and groups of letters (i.e. sequences of letters) into features is included in the handwriting model. After recognition has been performed, letters in the output text and the corresponding feature vectors used for recognition can be added to the lexicon. - Groups of commonly co-occurring letters and their corresponding feature vectors are concatenated and added to the lexicon. This is desirable since handwriting exhibits co-articulation effects (where the writing of a letter is influenced by the shape of surrounding letters) and commonly written letters (e.g. “qu”, “ed”, and “ing”) are likely to exhibit co-articulation. By storing the feature vectors of letter groups, a more accurate rendering of the query ink can be produced that considers the contextual effects of the stroke sequence.
- The lexicon should be able to store multiple feature vectors for each letter sequence. Since the same letters can be recognized many times, the lexicon should be able to select the feature vectors that best represents the letters. This can be done by storing all feature vector sequences output by the recognizer for each letter sequence, together with a count of the number of times that feature vector sequence has been seen for that letter sequence. The feature vector sequence with the highest count (i.e. the most frequently seen and thus most probable) for a letter sequence is then used during feature synthesis.
- The choice of which letter groups should be stored can be based on character transition statistics (for example, as derived from a text corpus) where character sequences with a high probability of occurrence are stored (e.g. the probability of “ing” is much greater than the probability of “inx”). Alternatively, all possible letter groupings can be stored after recognition, with some kind of culling procedure executed when the table becomes too large (for example, least-recently used).
- Further improvements can be achieved by explicitly modeling word endings in the lexicon. Many writing styles exhibit poorly written characters at the end of words. This is particularly apparent in handwritten word suffixes such as “ing”, “er”, and “ed”. To model this behavior, an end of word character is appended to the letter sequence (e.g. “ing#”) to indicate that this letter sequence models a word ending. During feature synthesis, these entries can only be used to complete a query word.
- To search a digital ink database, the text query is converted to a set of feature vectors by the feature synthesis procedure that uses the writer-specific handwriting model. These feature vectors are then used as the query term for searching the digital ink database. The ink database search can be performed using traditional ink matching techniques.
FIG. 3 describes this procedure. Atstep 40 text is input and thetext 41 provided to featuresynthesis step 42 which uses thehandwriting model 35 to generate features 43. Thefeatures 43 are used in anink searching step 44 of thedigital ink database 11. This produces ink matches 45. - To perform feature synthesis, a search of the lexicon is performed to locate a sequence of lexicon entries that contain letter sequences that can be combined to produce the query text. The stored feature vectors are concatenated to produce the query feature vector. However, there may be a number of different combinations of lexicon entries that can be used to create the query text. It is assumed that using the minimum number of entries possible to produce the text will most accurately model the contextual effects. For example, assume the following entries exists in the lexicon:
-
TABLE 1 Example lexicon b bo bor borr ed ow rr rowed - The word “borrowed” can then be created as (bo)(rr)(ow)(ed), (borr)(ow)(ed), or (bor)(rowed), with the final construction being most the desirable since it is composed from the fewest elements.
- For accurate results, the lexicon is expected to be very large, and the number of potential lexicon entry combinations for a word would be exponential. For queries containing long words, a complete enumeration of all permutations may not be practical. A modified A* search algorithm, see S. Russell and P. Norvig, Artificial Intelligence—A Modern Approach, Prentice Hall, 1995, can be used to search for a letter sequence s, where the path cost function g(s) is the number of lexicon entries used so far to create the text, and the estimated cost to the goal is:
-
- This heuristic states that at least one additional lexicon entry is required to complete the letter sequence if the sequence has fewer letters than the query word. Nodes in the search tree are sorted by g(s)+h(s) (with lower scores being superior) and nodes with the same score are ordered by the number of letters in the sequence (with higher being superior).
- Note that h(s) is a monotonic and admissible heuristic (i.e. it never overestimates the cost of reaching the goal) so the search is guaranteed to find the optimal solution and is optimally efficient (i.e. expands the fewest nodes possible to find the optimal solution). Proof of this result is given in R. Dechter and J. Pearl, “Generalized Best-First Search Strategies and the Optimality of A*”, Journal of the Association for Computing Machinery, 32(3), pp. 505-536, 1985.
- As an example of the above procedure, the search for the word “borrowed” is given below. Each row in the table represents a search node, with higher scoring nodes located at the top of the table:
-
s g(s) h(s) g(s) + h(s) # Letters (borr) 1 1 2 4 (bor) 1 1 2 3 (bo) 1 1 2 2 - The most promising node at the top of the table is expanded, resulting in:
-
s g(s) h(s) g(s) + h(s) # Letters (bor) 1 1 2 3 (bo) 1 1 2 2 (borr)(ow) 2 1 3 6 - Again, the most promising node is expanded:
-
s g(s) h(s) g(s) + h(s) # Letters (bor)(rowed) 2 0 2 8 (bo) 1 1 2 2 (borr)(ow) 2 1 3 6 - The top node is now a completed sequence, and no other node in the search can produce a better score, so this node is selected as the search result.
- IIA. Feature Synthesis without Handwriting Recognition
- Ink searching using feature synthesis can be performed without using a handwriting recognition system. The technique only requires the ability to build a lexicon of letter sequences and associated feature vectors to model the handwriting of the writer who authored the digital ink database.
- If handwriting recognition results are not available for modelling, a training procedure can be used to generate the writer-specific lexicon. To do this, a user provides a sample of their handwriting by copying a specified training text, which is then used to build the lexicon. The training procedure is not required to perform full handwriting recognition since the text represented by the handwriting is already known; rather, it simply needs to segment the input into characters and strokes, convert the strokes into features, and store the appropriate letter groups and associated feature vectors in the lexicon.
- The training text used to build the lexicon should be optimized to provide a balanced example set of individual letters and letter groups. That is, it should maximize the coverage of likely character unigrams, bigrams, and trigrams, see J. Pitrelli, J. Subrahmonia, M. Perrone, and K. Nathan, “Optimization of Training Texts for Writer-Dependent Handwriting Recognition”, Advances in Handwriting Recognition, World Scientific Publishing, 1999, with an emphasis on letters and letter sequences that are most likely to be encountered.
- It is desirable that both the handwriting recognition system and the ink-matching algorithm use the same feature representations, since the features used to search the digital ink database are derived from the results of the handwriting recognition.
- However, it is possible to use different feature sets for the recognition and search provided the recognition features can be transformed into the search features. Some feature sets can allow a transformation from recognition features to search features to be learned from a set of training data.
- Alternatively, many feature sets can allow the regeneration of an approximation of the digital ink from the recognition features, from which the second feature set can be extracted. That is, the text query is converted to a set of feature vectors using feature synthesis, and the inverse transformation of the feature-extraction process is applied to the features to convert them into digital ink from which the search features are extracted. Care must be taken to ensure that this procedure does not introduce artifacts in the digital ink that may affect the search feature extraction (e.g. discontinuities in the generated ink may cause problems for some feature extraction techniques).
- A person can search the digital ink database of another writer by using handwriting recognition to convert their ink input into text, and using feature synthesis to convert the recognized text into features for ink searching.
-
FIG. 4 depicts this situation, with writer B searching the digital ink database authored by writer A. Atstep 46 ink is received from writer B and the strokes 47 are passed to featureextraction step 48. Thefeatures 49 are extracted andrecognition step 50 uses alanguage model 51 andwriter B model 52 to generate correspondingtext 53. This is used to build a lexicon for writer B atlexicon builder step 54. Letter groups and features are sent back towriter B model 52 to improve/update the model. Thetext 53 then undergoesfeature synthesis step 42 and a similar process as that described with reference toFIG. 3 is then followed to retrieve ink matches 45 that have been authored by writer A. - The following example provides a more detailed outline of one embodiment of the present invention. This example is intended to be merely illustrative and not limiting of the scope of the present invention.
- This section describes in detail an implementation of ink database searching using feature synthesis. It is assumed that a number of preprocessing steps have been performed, including word and character segmentation, and baseline orientation normalization. Note that this is just one possible way to implement the technique; there are alternate methods available for each stage of the process. For example, there are many different segmentation schemes, feature sets, handwriting models, and recognition procedures that could also be used.
- The procedure for handwriting recognition and ink searching using feature synthesis requires a training phase and a recognition or searching phase. During the training phase, a set of training data is converted to stroke features that are clustered into stroke primitives and used to create a model of the handwriting. For recognition and ink searching, this model is used to decode input ink or synthesize features for ink searching. This process is depicted in
FIG. 5 . Note that the same preprocessing, normalization, segmentation, and feature extraction procedure is used for training, recognition and searching. - The handwriting recognition system maps input ink into a set of stroke code words that are used to search a dictionary for matching words.
FIG. 6 provides an overview of this process, with the individual steps described below in more detail. - The ink is sampled at a constant rate of 100 Hz. Research has shown that handwriting has a peak spectral density at around 5 Hz that declines to noise level at about 10 Hz, see H. Teulings and F. Maarse, “Digital Recording and Processing of Handwriting Movements”, Human Movement Science, 3, pp. 193-217, 1984. Thus, a low-pass filter with a cutoff at 10 Hz will remove the high-frequency noise without affecting the relevant spectral components of the handwriting signal.
- A low-pass filter conforming to the above specifications can be produced by circularizing the point coordinates, performing an FFT to remove the high frequency components, and recreating the signal using an inverse-FFT. However, a simple weighted averaging filter works as effectively. To smooth a sequences of points {p1 . . . pn):
-
- The filter width k and α smoothing coefficients are determined empirically.
- Many handwriting styles do not have a vertical principal axis of the drawn letters (i.e. the letters are drawn at a consistent slant). Removing handwriting slant is a normalization that can improve the recognition of handwritten letters. In handwriting, down-strokes are considered the most stable and consistent strokes, and thus are useful for detecting handwritten slant.
- To detect handwritten slant, a weighted-average direction of the down strokes in points {p1 . . . pn} is estimated:
-
- Angles α1 and α2 define which stroke segments are considered as parts of a down stroke and are empirically set at 40° and 140° respectively (with 90° representing a vertical line). If the estimated slant deviates more than a certain threshold from the vertical, slant is removed using a shear transformation:
-
- where ymin and ymax represents the top and bottom of the bounding rectangle of the ink.
- Zone estimation is used to normalize the height of the input ink. English letters exhibit three zones—the middle zone (corresponding to the height of letters such as ‘a’, ‘c’, ‘e’, etc.), and the upper and lower zones that contain the ascenders and descenders in letters such as ‘b’, ‘d’, ‘g’, and ‘j’.
- Zone estimation is performed using a horizontal histogram of ink density. That is, the number of ink crossings is determined for an equally spaced series of horizontal lines passing through the bounding rectangle of the ink. The central peak of the histogram is found, as are the two points on either side of the histogram where the ink density drops below a certain fraction of the central peak height. These two points are selected as the upper and lower bounds of the middle zone. The upper and lower zones are defined as the space between the middle zone and the vertical extremum of the bounding rectangle.
- The ink is segmented into sub-strokes at extrema in the vertical direction (i.e. at local maxima and minima of the Y coordinates). For segmentation to occur, the lengths of the two sub-strokes produced by splitting the stroke at the selected segmentation point must exceed a pre-calculated minimum distance (set at half the height of the estimated middle zone).
- The segmented sub-strokes are then re-sampled to contain a constant n number of points located equidistantly along the stroke trajectory. A feature vector is then created for the sub-stroke by normalizing the coordinates:
-
- where
-
- xmin=X minimum of sub-stroke bounding rectangle
- ymiddle=Y coordinate of the top of the middle zone
- h=the height of the middle zone (i.e. ybase−ymiddle)
- The feature vector is then created from the normalized coordinates as fi={x′l,y′l, . . . , x′n,y′n}.
- The resulting vectors describe the sub-strokes using a large number of highly correlated features (clearly, the coordinates of point pi are dependent on point pi-1 and so on). To lower the dimensionality of the vectors to m (where m<2n), the Karhunen-Loeve transform (see Principal Component Analysis—R. Duda, P. Hart, and D. Stork, Pattern Classification, Second Edition, John Wiley & Sons, Inc., pp. 569-570, 2001) is used. This procedure projects the higher-dimensionality features into a lower dimension using linear mapping that is optimal in a least-squares sense.
- To do this, the covariance (autocorrelation) matrix for the set of all training feature vectors X={f1, . . . , fn} is calculated using:
-
- The eigenvectors and eigenvalues for this matrix are found (using the tridiagonal QL implicit algorithm, see W. Press, B. Flannery, S. Teukolsky, and W. Vetterling, W. T., Numerical Recipes in C, Cambridge: Cambridge University Press, 1988), and the eigenvectors corresponding to the largest n eigenvalues are used to form the PCA matrix Z. The feature vectors are then multiplied by this matrix to transform the features into a new feature space with an orthogonal uncorrelated axis:
-
f i ′=f i T Z - The transformed feature vectors are then clustered using a Kohonen Self-Organizing Feature Map (SOFM), see T. Kohonen, “Self-Organized Formation of Topologically Correct Feature Maps”, Biological Cybernetics, 43, pp. 59-69, 1982. This technique uses an unsupervised learning procedure to cluster the input vectors such that distance and proximity relationships between the vectors is maintained as far as possible. The SOFM used has a 2-dimensional structure, with visually similar code words (i.e. clusters) located near each other. As a result, the distance between two code words can be easily calculated using some distance measure (e.g. Euclidean distance) between the code word values.
- SOFM training is performed iteratively using a simple two-layer neural network that is initialized with random weights. The best matching output neuron for a normalized input training vector x is found using minimum Euclidean distance:
-
- where wi represents the weight vector of output node i. The weights of the node with the highest activation value and those nodes surrounding it (as determined by a neighborhood function Λ) are updated using:
-
w i(t+1)=w i(t)+η(t)[x−w i(t)] - where η is a learning rate function and η and Λ are typically varied over time t. Training continues until there is no noticeable change to the neuron weights during an iteration of the training set.
- To convert a sequence of sub-strokes to a code word vector, each sub-stroke feature vector is quantized using the trained SOFM and appended to the code word vector. A feature vector is quantized into a code word by selecting the largest activation value for the output neurons of the SOFM codebook:
-
- The handwriting model stores a mapping of stroke code word vectors to letters (for text recognition) and the reverse mapping of letter groups to code word vectors (for feature synthesis). To build the handwriting model, each letter in the training data is converted to a code word vector that is stored in a table, along with the corresponding letter.
- Note that a specific code word sequence may map to a number of letters (e.g. poorly-drawn ‘u’ may map to the same feature vectors as a ‘v’) and individual letters may be mapped to by a number of code word vectors. By maintaining a count of the number of times the code word vector represented a specific letter, the probability that the vector represents the letter xi can be calculated given n lexicon entries for letter xi:
-
- where ci is the count of the number of times the code word vector has been encountered representing the letter xi. The following is an example entry from the table that represents the hypothetical code word vector {3,4}:
-
TABLE 2 Example code word vector entry xi cI P(xi) ‘u’ 120 0.54 ‘v’ 91 0.41 ‘r’ 12 0.05 - This table indicates that if the code word sequence {3,4} is encountered in the input, there is a probability of 0.54 that it represents the letter ‘u’, a probability 0.41 that it represents an ‘v’, and a probability of 0.05 that it represents an ‘r’.
- The reverse mapping table is produced in a similar way, storing code word vectors associated with letters and groups of letters.
- To perform handwriting recognition, the input ink is processed as described above, and the resulting code word vectors are used to search the handwriting model to produce letter hypotheses. A best-first search strategy is used to search a dictionary for the most likely word given the derived letter probabilities. The process is depicted below in
FIG. 7 . - Ink searching is performed by using the handwriting model generated during handwriting recognition to map the input query text into a sequence of code words. This code word vector is used to search the digital ink database using an elastic matching technique. A full description of a similar ink-matching technique is given in D. Lopresti and A. Tomkins, “Temporal-Domain Matching of Hand-Drawn Pictorial Queries”, Handwriting and Drawing Research: Basic and Applied Issues, IOS Press, pp. 387-401, 1996. The resulting queries are then ordered by similarity and presented to the user.
- Thus, there has been provided in accordance with the present invention, a method and apparatus for searching a digital ink database using a text query which satisfies the advantages set forth above.
- The invention may also be said broadly to consist in the parts, elements and features referred to or indicated in the specification of the application, individually or collectively, in any or all combinations of two or more of said parts, elements or features, and where specific integers are mentioned herein which have known equivalents in the art to which the invention relates, such known equivalents are deemed to be incorporated herein as if individually set forth.
- Although the preferred embodiment has been described in detail, it should be understood that various changes, substitutions, and alterations can be made herein by one of ordinary skill in the art without departing from the scope of the present invention as hereinbefore described and as hereinafter claimed.
Claims (6)
1. A method of searching a digital ink database, the digital ink database being associated with a predetermined author, the method comprising the steps of:
receiving a computer text query from an input device;
mapping by a processing system the computer text query to a set of feature vectors using a handwriting model of the predetermined author, the set of feature vectors approximating features that would have been extracted had the predetermined author written the computer query text by hand; and
using the set of feature vectors to search the digital ink database.
2. The method according to claim 1 wherein the input device is a computer.
3. The method according to claim 1 wherein the input device is a tablet for receiving hand written input, and the receiving step comprises the sub-step of converting the hand written input to computer text through handwriting recognition.
4. The method according to claim 3 wherein the hand written input is received from an author other that the predetermined author, and the set of feature vectors extracted during the handwriting recognition is different from the set of feature vectors used when searching the digital ink database.
5. The method according to claim 1 wherein the input device receives a spoken voice input, and the receiving step comprises the sub-step of converting the spoken voice input to computer text through voice recognition.
6. An apparatus for searching a digital ink database, the digital ink database being associated with a predetermined author, the apparatus comprising:
an input device for receiving a computer text query;
a processing system for mapping by the computer text query to a set of feature vectors using a handwriting model of the predetermined author, the set of feature vectors approximating features that would have been extracted had the predetermined author written the computer query text by hand, and for using the set of feature vectors to search the digital ink database; and
an output device for outputting results of the search of the digital ink database.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US13/482,828 US20120239397A1 (en) | 2001-10-15 | 2012-05-29 | Digital Ink Database Searching Using Handwriting Feature Synthesis |
Applications Claiming Priority (5)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
AUPR8243A AUPR824301A0 (en) | 2001-10-15 | 2001-10-15 | Methods and systems (npw001) |
AUPR8243 | 2001-10-15 | ||
US10/492,169 US8219908B2 (en) | 2001-10-15 | 2002-10-15 | Digital ink database searching using handwriting feature synthesis |
PCT/AU2002/001395 WO2003034276A1 (en) | 2001-10-15 | 2002-10-15 | Digital ink database searching using handwriting feature synthesis |
US13/482,828 US20120239397A1 (en) | 2001-10-15 | 2012-05-29 | Digital Ink Database Searching Using Handwriting Feature Synthesis |
Related Parent Applications (2)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US10/492,169 Continuation US8219908B2 (en) | 2001-10-15 | 2002-10-15 | Digital ink database searching using handwriting feature synthesis |
PCT/AU2002/001395 Continuation WO2003034276A1 (en) | 2001-10-15 | 2002-10-15 | Digital ink database searching using handwriting feature synthesis |
Publications (1)
Publication Number | Publication Date |
---|---|
US20120239397A1 true US20120239397A1 (en) | 2012-09-20 |
Family
ID=3832068
Family Applications (2)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US10/492,169 Expired - Fee Related US8219908B2 (en) | 2001-10-15 | 2002-10-15 | Digital ink database searching using handwriting feature synthesis |
US13/482,828 Abandoned US20120239397A1 (en) | 2001-10-15 | 2012-05-29 | Digital Ink Database Searching Using Handwriting Feature Synthesis |
Family Applications Before (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US10/492,169 Expired - Fee Related US8219908B2 (en) | 2001-10-15 | 2002-10-15 | Digital ink database searching using handwriting feature synthesis |
Country Status (12)
Country | Link |
---|---|
US (2) | US8219908B2 (en) |
EP (1) | EP1444614B1 (en) |
JP (2) | JP2005505861A (en) |
KR (1) | KR100717637B1 (en) |
CN (1) | CN1701323B (en) |
AT (1) | ATE358851T1 (en) |
AU (1) | AUPR824301A0 (en) |
CA (1) | CA2463236C (en) |
DE (1) | DE60219342D1 (en) |
IL (1) | IL161381A0 (en) |
WO (1) | WO2003034276A1 (en) |
ZA (1) | ZA200402928B (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8640046B1 (en) * | 2012-10-23 | 2014-01-28 | Google Inc. | Jump scrolling |
CN107562716A (en) * | 2017-07-18 | 2018-01-09 | 阿里巴巴集团控股有限公司 | Term vector processing method, device and electronic equipment |
CN107562715A (en) * | 2017-07-18 | 2018-01-09 | 阿里巴巴集团控股有限公司 | Term vector processing method, device and electronic equipment |
US9880990B2 (en) | 2013-03-14 | 2018-01-30 | Microsoft Technology Licensing, Llc | Leveraging previous instances of handwriting for handwriting beautification and other applications |
CN107844472A (en) * | 2017-07-18 | 2018-03-27 | 阿里巴巴集团控股有限公司 | Term vector processing method, device and electronic equipment |
US10380483B2 (en) * | 2015-01-19 | 2019-08-13 | Samsung Electronics Co., Ltd. | Method and apparatus for training language model, and method and apparatus for recognizing language |
US20220291827A1 (en) * | 2021-03-09 | 2022-09-15 | Societe Bic | Handwriting feedback |
Families Citing this family (50)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
AU2002952483A0 (en) | 2002-11-05 | 2002-11-21 | Silverbrook Research Pty Ltd | Methods and Systems (NPW009) |
CA2529037A1 (en) * | 2003-08-15 | 2005-02-24 | Silverbrook Research Pty Ltd | Natural language recognition using distributed processing |
AU2004265699B2 (en) * | 2003-08-15 | 2007-08-09 | Silverbrook Research Pty Ltd | Improving accuracy in searching digital ink |
FR2880709B1 (en) * | 2005-01-11 | 2014-04-25 | Vision Objects | METHOD OF SEARCHING, RECOGNIZING AND LOCATING INK, DEVICE, PROGRAM AND LANGUAGE CORRESPONDING |
US7593908B2 (en) | 2005-06-27 | 2009-09-22 | Microsoft Corporation | Training with heterogeneous data |
US7646940B2 (en) * | 2006-04-04 | 2010-01-12 | Microsoft Corporation | Robust indexing and retrieval of electronic ink |
US8094939B2 (en) * | 2007-06-26 | 2012-01-10 | Microsoft Corporation | Digital ink-based search |
US8041120B2 (en) | 2007-06-26 | 2011-10-18 | Microsoft Corporation | Unified digital ink recognition |
US8315482B2 (en) * | 2007-06-26 | 2012-11-20 | Microsoft Corporation | Integrated platform for user input of digital ink |
CN101482920B (en) * | 2008-12-30 | 2010-12-22 | 广东国笔科技股份有限公司 | Hand-written character recognition method and system |
CN102117122B (en) * | 2009-12-30 | 2014-08-13 | 浪潮乐金数字移动通信有限公司 | English input method and mobile communication terminal |
CN101901355A (en) * | 2010-06-29 | 2010-12-01 | 北京捷通华声语音技术有限公司 | Character recognition method and device based on maximum entropy |
CN102005033B (en) * | 2010-11-16 | 2012-07-04 | 中国科学院遥感应用研究所 | Method for suppressing noise by image smoothing |
CN102033928B (en) * | 2010-12-15 | 2013-01-23 | 北京壹人壹本信息科技有限公司 | Handwritten original handwriting recording and querying system and electronic device |
KR101830787B1 (en) | 2011-07-12 | 2018-02-21 | 삼성전자 주식회사 | Method and apparatus for searching hand written memo data |
US10430066B2 (en) * | 2011-12-06 | 2019-10-01 | Nri R&D Patent Licensing, Llc | Gesteme (gesture primitive) recognition for advanced touch user interfaces |
KR101898202B1 (en) * | 2012-02-09 | 2018-09-12 | 삼성전자주식회사 | Apparatus and method for guiding writing input for recognation of writing |
US8935283B2 (en) | 2012-04-11 | 2015-01-13 | Blackberry Limited | Systems and methods for searching for analog notations and annotations |
JP5717691B2 (en) * | 2012-05-28 | 2015-05-13 | 株式会社東芝 | Handwritten character search device, method and program |
JP5330576B1 (en) * | 2012-07-09 | 2013-10-30 | 株式会社東芝 | Information processing apparatus and handwriting search method |
KR101997447B1 (en) * | 2012-12-10 | 2019-07-08 | 엘지전자 주식회사 | Mobile terminal and controlling method thereof |
US8943092B2 (en) | 2013-03-04 | 2015-01-27 | Microsoft Corporation | Digital ink based contextual search |
US9558400B2 (en) | 2013-03-07 | 2017-01-31 | Ricoh Company, Ltd. | Search by stroke |
WO2014147719A1 (en) * | 2013-03-18 | 2014-09-25 | 株式会社 東芝 | Electronic device, and method for processing handwritten document |
JP6125333B2 (en) * | 2013-05-31 | 2017-05-10 | 株式会社東芝 | Search device, method and program |
US20140363082A1 (en) * | 2013-06-09 | 2014-12-11 | Apple Inc. | Integrating stroke-distribution information into spatial feature extraction for automatic handwriting recognition |
US9495620B2 (en) | 2013-06-09 | 2016-11-15 | Apple Inc. | Multi-script handwriting recognition using a universal recognizer |
US9465985B2 (en) | 2013-06-09 | 2016-10-11 | Apple Inc. | Managing real-time handwriting recognition |
US9201592B2 (en) * | 2013-08-09 | 2015-12-01 | Blackberry Limited | Methods and devices for providing intelligent predictive input for handwritten text |
KR101488670B1 (en) * | 2014-01-29 | 2015-02-04 | 주식회사 디오텍 | System and method for cross-searching large-scale documents including text and digital ink |
JP6270565B2 (en) * | 2014-03-18 | 2018-01-31 | 株式会社東芝 | Electronic apparatus and method |
US10997757B1 (en) * | 2014-06-17 | 2021-05-04 | FlipScript, Inc. | Method of automated typographical character modification based on neighboring characters |
CN105809100B (en) * | 2014-12-31 | 2019-04-19 | 中国刑事警察学院 | Person's handwriting and the signature quantization total area method of test sensitivity feature |
FR3038100A1 (en) * | 2015-06-26 | 2016-12-30 | Orange | METHOD OF RECOGNIZING HANDWRITING ON A PHYSICAL SURFACE |
KR101669821B1 (en) * | 2015-11-09 | 2016-10-27 | 주식회사 셀바스에이아이 | Method and apparatus for generating hand-writing data base and computer program stored in computer readable medium therefor |
KR101667910B1 (en) * | 2015-11-09 | 2016-10-20 | 주식회사 셀바스에이아이 | Method and apparatus for generating digital artifical hand-writing data and computer program stored in computer readable medium therefor |
US10228775B2 (en) * | 2016-01-22 | 2019-03-12 | Microsoft Technology Licensing, Llc | Cross application digital ink repository |
CN107292213B (en) * | 2016-03-30 | 2020-04-14 | 中国刑事警察学院 | Handwriting quantitative inspection and identification method |
DK179329B1 (en) | 2016-06-12 | 2018-05-07 | Apple Inc | Handwriting keyboard for monitors |
KR102445371B1 (en) * | 2017-12-11 | 2022-09-20 | 현대자동차주식회사 | Character recognition method of a handwriting input device, handwriting input device, and vehicle having handwriting input device |
US10514841B2 (en) * | 2018-04-30 | 2019-12-24 | Microsoft Technology Licensing, Llc | Multi-layered ink object |
CN109002849B (en) * | 2018-07-05 | 2022-05-17 | 百度在线网络技术(北京)有限公司 | Method and device for identifying development stage of object |
US11194467B2 (en) | 2019-06-01 | 2021-12-07 | Apple Inc. | Keyboard management user interfaces |
CN111310543B (en) * | 2019-12-04 | 2023-05-30 | 湖北工业大学 | Method for extracting and authenticating stroke-extracting continuous stroke characteristics in online handwriting authentication |
CN113128296B (en) * | 2019-12-31 | 2023-05-09 | 重庆傲雄在线信息技术有限公司 | Electronic handwriting signature fuzzy label recognition system |
CN111242840A (en) * | 2020-01-15 | 2020-06-05 | 上海眼控科技股份有限公司 | Handwritten character generation method, apparatus, computer device and storage medium |
CN113378609B (en) * | 2020-03-10 | 2023-07-21 | 中国移动通信集团辽宁有限公司 | Agent proxy signature identification method and device |
CN112241686A (en) * | 2020-09-16 | 2021-01-19 | 四川天翼网络服务有限公司 | Trajectory comparison matching method and system based on feature vectors |
WO2022087847A1 (en) * | 2020-10-27 | 2022-05-05 | 京东方科技集团股份有限公司 | Handwritten text recognition method, apparatus and system, handwritten text search method and system, and computer-readable storage medium |
CN113610065B (en) * | 2021-10-09 | 2022-05-20 | 北京世纪好未来教育科技有限公司 | Handwriting recognition method and device |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5933823A (en) * | 1996-03-01 | 1999-08-03 | Ricoh Company Limited | Image database browsing and query using texture analysis |
US6480587B1 (en) * | 1996-12-16 | 2002-11-12 | Sanjay K. Rao | Intelligent keyboard system |
US7092870B1 (en) * | 2000-09-15 | 2006-08-15 | International Business Machines Corporation | System and method for managing a textual archive using semantic units |
Family Cites Families (43)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4864618A (en) * | 1986-11-26 | 1989-09-05 | Wright Technologies, L.P. | Automated transaction system with modular printhead having print authentication feature |
US5051736A (en) * | 1989-06-28 | 1991-09-24 | International Business Machines Corporation | Optical stylus and passive digitizing tablet data input system |
US5063600A (en) * | 1990-05-14 | 1991-11-05 | Norwood Donald D | Hybrid information management system for handwriting and text |
US5220649A (en) * | 1991-03-20 | 1993-06-15 | Forcier Mitchell D | Script/binary-encoded-character processing method and system with moving space insertion mode |
US5852434A (en) * | 1992-04-03 | 1998-12-22 | Sekendur; Oral F. | Absolute optical position determination |
US5477012A (en) * | 1992-04-03 | 1995-12-19 | Sekendur; Oral F. | Optical position determination |
US5583543A (en) * | 1992-11-05 | 1996-12-10 | Sharp Kabushiki Kaisha | Pen input processing apparatus |
US5392363A (en) * | 1992-11-13 | 1995-02-21 | International Business Machines Corporation | On-line connected handwritten word recognition by a probabilistic method |
IT1265673B1 (en) * | 1993-03-22 | 1996-11-22 | Texas Instruments Italia Spa | APPARATUS AND PROCEDURE FOR THE RECOGNITION OF MANUSCRIPT CHARACTERS. |
JP2933801B2 (en) * | 1993-06-11 | 1999-08-16 | 富士通株式会社 | Method and apparatus for cutting out characters |
US5454046A (en) * | 1993-09-17 | 1995-09-26 | Penkey Corporation | Universal symbolic handwriting recognition system |
US5687254A (en) * | 1994-06-06 | 1997-11-11 | Xerox Corporation | Searching and Matching unrecognized handwriting |
US5652412A (en) * | 1994-07-11 | 1997-07-29 | Sia Technology Corp. | Pen and paper information recording system |
US5666438A (en) * | 1994-07-29 | 1997-09-09 | Apple Computer, Inc. | Method and apparatus for recognizing handwriting of different users of a pen-based computer system |
US5768423A (en) * | 1994-09-02 | 1998-06-16 | Panasonic Technologies Inc. | Trie structure based method and apparatus for indexing and searching handwritten databases with dynamic search sequencing |
US5768417A (en) * | 1994-09-09 | 1998-06-16 | Motorola, Inc. | Method and system for velocity-based handwriting recognition |
GB2307085B (en) * | 1994-09-09 | 1998-09-23 | Motorola Inc | Method and system for recognizing a boundary between characters in handwritten text |
WO1996008787A1 (en) * | 1994-09-14 | 1996-03-21 | Apple Computer, Inc. | System and method for automatic subcharacter unit and lexicon generation for handwriting recognition |
US5661506A (en) * | 1994-11-10 | 1997-08-26 | Sia Technology Corporation | Pen and paper information recording system using an imaging pen |
US5757959A (en) * | 1995-04-05 | 1998-05-26 | Panasonic Technologies, Inc. | System and method for handwriting matching using edit distance computation in a systolic array processor |
US6081261A (en) | 1995-11-01 | 2000-06-27 | Ricoh Corporation | Manual entry interactive paper and electronic document handling and processing system |
US5832474A (en) * | 1996-02-26 | 1998-11-03 | Matsushita Electric Industrial Co., Ltd. | Document search and retrieval system with partial match searching of user-drawn annotations |
US5692073A (en) * | 1996-05-03 | 1997-11-25 | Xerox Corporation | Formless forms and paper web using a reference-based mark extraction technique |
DE69724405T2 (en) * | 1996-10-04 | 2004-05-06 | Koninklijke Philips Electronics N.V. | METHOD AND APPARATUS FOR ONLINE HANDWRITING RECOGNITION BASED ON CHARACTER VECTORS USING AGGLOMERED OBSERVATIONS FROM TIME-SEQUENT SEQUENCES |
US6275611B1 (en) * | 1996-10-17 | 2001-08-14 | Motorola, Inc. | Handwriting recognition device, method and alphabet, with strokes grouped into stroke sub-structures |
US5970455A (en) * | 1997-03-20 | 1999-10-19 | Xerox Corporation | System for capturing and retrieving audio data and corresponding hand-written notes |
JP3020887B2 (en) | 1997-04-14 | 2000-03-15 | 株式会社エイ・ティ・アール知能映像通信研究所 | Database storage method, database search method, and database device |
US5953451A (en) * | 1997-06-19 | 1999-09-14 | Xerox Corporation | Method of indexing words in handwritten document images using image hash tables |
US6518950B1 (en) * | 1997-10-07 | 2003-02-11 | Interval Research Corporation | Methods and systems for providing human/computer interfaces |
WO1999018556A2 (en) * | 1997-10-08 | 1999-04-15 | Koninklijke Philips Electronics N.V. | Vocabulary and/or language model training |
WO1999050751A1 (en) | 1998-04-01 | 1999-10-07 | Xerox Corporation | Routing document identifiers |
US6157731A (en) * | 1998-07-01 | 2000-12-05 | Lucent Technologies Inc. | Signature verification method using hidden markov models |
US6964374B1 (en) * | 1998-10-02 | 2005-11-15 | Lucent Technologies Inc. | Retrieval and manipulation of electronically stored information via pointers embedded in the associated printed material |
JP2001148123A (en) | 1999-11-22 | 2001-05-29 | Sony Corp | Optical disk, recording and/or reproducing device and method |
US6355889B1 (en) * | 2000-06-28 | 2002-03-12 | International Business Machines Corporation | Method and apparatus for linking electronic ink to electronic personal information systems |
US7155061B2 (en) * | 2000-08-22 | 2006-12-26 | Microsoft Corporation | Method and system for searching for words and phrases in active and stored ink word documents |
US6785417B1 (en) * | 2000-08-22 | 2004-08-31 | Microsoft Corp | Method and system for searching for words in ink word documents |
US7336827B2 (en) * | 2000-11-08 | 2008-02-26 | New York University | System, process and software arrangement for recognizing handwritten characters |
US20020107885A1 (en) * | 2001-02-01 | 2002-08-08 | Advanced Digital Systems, Inc. | System, computer program product, and method for capturing and processing form data |
US7627596B2 (en) * | 2001-02-22 | 2009-12-01 | International Business Machines Corporation | Retrieving handwritten documents using multiple document recognizers and techniques allowing both typed and handwritten queries |
US7284191B2 (en) * | 2001-08-13 | 2007-10-16 | Xerox Corporation | Meta-document management system with document identifiers |
US7188313B2 (en) * | 2001-10-03 | 2007-03-06 | Hallmark Cards, Incorporated | Context sensitive font generation |
US7523235B2 (en) * | 2003-06-11 | 2009-04-21 | Lsi Corporation | Serial Advanced Technology Attachment (SATA) switch |
-
2001
- 2001-10-15 AU AUPR8243A patent/AUPR824301A0/en not_active Abandoned
-
2002
- 2002-10-15 CA CA2463236A patent/CA2463236C/en not_active Expired - Fee Related
- 2002-10-15 JP JP2003536935A patent/JP2005505861A/en active Pending
- 2002-10-15 KR KR1020047005557A patent/KR100717637B1/en not_active IP Right Cessation
- 2002-10-15 IL IL16138102A patent/IL161381A0/en unknown
- 2002-10-15 DE DE60219342T patent/DE60219342D1/en not_active Expired - Lifetime
- 2002-10-15 WO PCT/AU2002/001395 patent/WO2003034276A1/en active IP Right Grant
- 2002-10-15 US US10/492,169 patent/US8219908B2/en not_active Expired - Fee Related
- 2002-10-15 AT AT02801227T patent/ATE358851T1/en not_active IP Right Cessation
- 2002-10-15 EP EP02801227A patent/EP1444614B1/en not_active Expired - Lifetime
- 2002-10-15 CN CN028204727A patent/CN1701323B/en not_active Expired - Fee Related
-
2004
- 2004-04-19 ZA ZA200402928A patent/ZA200402928B/en unknown
-
2008
- 2008-07-04 JP JP2008175821A patent/JP2008293519A/en active Pending
-
2012
- 2012-05-29 US US13/482,828 patent/US20120239397A1/en not_active Abandoned
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5933823A (en) * | 1996-03-01 | 1999-08-03 | Ricoh Company Limited | Image database browsing and query using texture analysis |
US6480587B1 (en) * | 1996-12-16 | 2002-11-12 | Sanjay K. Rao | Intelligent keyboard system |
US7092870B1 (en) * | 2000-09-15 | 2006-08-15 | International Business Machines Corporation | System and method for managing a textual archive using semantic units |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8640046B1 (en) * | 2012-10-23 | 2014-01-28 | Google Inc. | Jump scrolling |
US9880990B2 (en) | 2013-03-14 | 2018-01-30 | Microsoft Technology Licensing, Llc | Leveraging previous instances of handwriting for handwriting beautification and other applications |
US10380483B2 (en) * | 2015-01-19 | 2019-08-13 | Samsung Electronics Co., Ltd. | Method and apparatus for training language model, and method and apparatus for recognizing language |
CN107562716A (en) * | 2017-07-18 | 2018-01-09 | 阿里巴巴集团控股有限公司 | Term vector processing method, device and electronic equipment |
CN107562715A (en) * | 2017-07-18 | 2018-01-09 | 阿里巴巴集团控股有限公司 | Term vector processing method, device and electronic equipment |
CN107844472A (en) * | 2017-07-18 | 2018-03-27 | 阿里巴巴集团控股有限公司 | Term vector processing method, device and electronic equipment |
US20220291827A1 (en) * | 2021-03-09 | 2022-09-15 | Societe Bic | Handwriting feedback |
US11675495B2 (en) * | 2021-03-09 | 2023-06-13 | SOCIéTé BIC | Handwriting feedback |
Also Published As
Publication number | Publication date |
---|---|
CA2463236A1 (en) | 2003-04-24 |
CN1701323B (en) | 2010-06-23 |
KR20050036863A (en) | 2005-04-20 |
AUPR824301A0 (en) | 2001-11-08 |
DE60219342D1 (en) | 2007-05-16 |
IL161381A0 (en) | 2004-09-27 |
ATE358851T1 (en) | 2007-04-15 |
US20050222848A1 (en) | 2005-10-06 |
KR100717637B1 (en) | 2007-05-15 |
WO2003034276A1 (en) | 2003-04-24 |
JP2005505861A (en) | 2005-02-24 |
EP1444614B1 (en) | 2007-04-04 |
US8219908B2 (en) | 2012-07-10 |
CN1701323A (en) | 2005-11-23 |
CA2463236C (en) | 2010-09-21 |
EP1444614A4 (en) | 2005-12-28 |
EP1444614A1 (en) | 2004-08-11 |
ZA200402928B (en) | 2005-02-23 |
JP2008293519A (en) | 2008-12-04 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US8219908B2 (en) | Digital ink database searching using handwriting feature synthesis | |
CN106776562B (en) | Keyword extraction method and extraction system | |
Hu et al. | Writer independent on-line handwriting recognition using an HMM approach | |
US7925663B2 (en) | Searching an electronic filing system using a handwritten search query and a text search query | |
Lladós et al. | On the influence of word representations for handwritten word spotting in historical documents | |
US8077973B2 (en) | Handwritten word recognition based on geometric decomposition | |
Sun et al. | Deep LSTM networks for online Chinese handwriting recognition | |
CN114254653A (en) | Scientific and technological project text semantic extraction and representation analysis method | |
Kaur et al. | A comprehensive survey on word recognition for non-Indic and Indic scripts | |
US20230297629A1 (en) | Apparatus and method of performance matching | |
ZA200200661B (en) | Method and apparatus for determining the cell border for handover operation. | |
US11854537B2 (en) | Systems and methods for parsing and correlating solicitation video content | |
US20230289396A1 (en) | Apparatuses and methods for linking posting data | |
US11538462B1 (en) | Apparatuses and methods for querying and transcribing video resumes | |
Chan et al. | Online text independent writer identification using character prototypes distribution | |
CN118152570A (en) | Intelligent text classification method | |
Calvo-Zaragoza et al. | Recognition of pen-based music notation with finite-state machines | |
CN113076758A (en) | Task-oriented dialog-oriented multi-domain request type intention identification method | |
US20230298571A1 (en) | Apparatuses and methods for querying and transcribing video resumes | |
AU2002333066B2 (en) | Digital ink database searching using handwriting feature synthesis | |
AU2002333066A1 (en) | Digital ink database searching using handwriting feature synthesis | |
ZA200507773B (en) | Method of searching handwritten text databases using a lexicon | |
Wu | Establishment of Japanese Continuous Speech Recognition System Based on Association Rules Mining | |
Kaur et al. | Online handwritten text recognition for indian scripts | |
Baiju | Pattern primitive based malayalam handwritten character recognition studies for real-time applications |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AS | Assignment |
Owner name: SILVERBROOK RESEARCH PTY LTD, AUSTRALIA Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:NAPPER, JONATHON LEIGH;LAPSTUN, PAUL;REEL/FRAME:028283/0183 Effective date: 20040326 |
|
STCB | Information on status: application discontinuation |
Free format text: ABANDONED -- FAILURE TO RESPOND TO AN OFFICE ACTION |