Cursive character handwriting recognition system and method
First Claim
1. A method for recognizing unconstrained cursive handwritten words, comprising:
- processing an image of a handwritten word of one or more characters, the processing step including segmenting each imaged word into a set of one or more segments and determining a sequence of segments using an over-segmentation-relabeling algorithm;
extracting feature information of one segment or a combination of several consecutive segments;
repeating said extracting step until feature information from all segments or combinations thereof have been extracted; and
classifying the imaged word as having a string of one or more characters using the extracted feature information.
1 Assignment
0 Petitions
Accused Products
Abstract
A cursive character handwriting recognition system includes image processing means for processing an image of a handwritten word of one or more characters and classification means for determining an optimal string of one or more characters as composing the imaged word. The processing means segments the characters such that each character is made up of one or more segments and determines a sequence of the segments using an over-segmentation-relabeling algorithm. The system also includes feature extraction means for deriving a feature vector to represent feature information of one segment or a combination of several consecutive segments. The over-segmentation-relabeling algorithm places certain segments considered as diacritics or small segments so as to immediately precede or follow a segment of the associated main character body. Additionally, the system also includes classification means that processes each string of segments and outputs a number of optimal strings which could be matched against a given lexicon.
71 Citations
29 Claims
-
1. A method for recognizing unconstrained cursive handwritten words, comprising:
-
processing an image of a handwritten word of one or more characters, the processing step including segmenting each imaged word into a set of one or more segments and determining a sequence of segments using an over-segmentation-relabeling algorithm; extracting feature information of one segment or a combination of several consecutive segments; repeating said extracting step until feature information from all segments or combinations thereof have been extracted; and classifying the imaged word as having a string of one or more characters using the extracted feature information. - View Dependent Claims (2, 3, 4)
-
-
5. An unconstrained cursive character handwritten word recognition system, comprising:
-
image processing means for processing an image of a handwritten word of one or more characters, wherein the processing of the imaged word includes segmenting the imaged word into a finite number of segments and determining a sequence of the segments using an over-segmentation-relabeling algorithm, wherein each character includes one or more segments; feature extraction means for deriving a feature vector to represent feature information of one segment or a combination of several consecutive segments; and classification means for determining an optimal string of one or more characters as composing the imaged word. - View Dependent Claims (6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20)
-
-
21. A method for training an unconstrained cursive character handwritten word recognition system, comprising:
-
processing a corpus of handwritten word images, each imaged word having one or more characters, the processing step including segmenting each imaged word into a set of one or more segments and determining a sequence of the segments using an over-segmentation-relabeling algorithm; extracting feature information of individual characters of the imaged words; estimating symbol probability parameters associated with each distinct character so as to allow a statistical measure that given feature information is indicative of a distinct character; and estimating state duration probabilities associated with each distinct character, wherein a state duration probability of a given distinct character represents a probability that a segmented image of the given character will have a duration of a defined number of segments. - View Dependent Claims (22, 23)
-
-
24. A method for determining a sequence of segments of a segmented image of a cursive written word processed in a word recognition system, comprising:
-
finding the number of segments, wherein the finding step includes locating the first segment and the last segment in the imaged word; and determining the sequence of segments using an over-segmentation-relabeling algorithm, wherein the over-segmentation-relabeling algorithm includes; characterizing segments as either situated segments or unsituated segments, wherein situated segments include the first and last segments and segments having an X-coordinate or Y-coordinate coverage that exceeds a threshold value, small segments that are cursively connected to segments on each side, and wherein unsituated segments are segments not characterized as situated segments; and placing each unsituated segment having a situated segment above or below so as to either immediately precede or follow the situated segment in the sequence of segments. - View Dependent Claims (25, 26, 27, 28, 29)
-
Specification