Title :
Chaincode contour processing for handwritten word recognition
Author :
Madhvanath, S. ; Kim, G. ; Govindaraju, V.
Author_Institution :
IBM Almaden Res. Center, San Jose, CA, USA
fDate :
9/1/1999 12:00:00 AM
Abstract :
Contour representations of binary images of handwritten words afford considerable reduction in storage requirements while providing lossless representation. On the other hand, the one-dimensional nature of contours presents interesting challenges for processing images for handwritten word recognition. Our experiments indicate that significant gains are to be realized in both speed and recognition accuracy by using a contour representation in handwriting applications
Keywords :
feature extraction; handwritten character recognition; image segmentation; binary images; chaincode contour processing; contour representations; handwritten word recognition; lossless representation; recognition accuracy; recognition speed; storage requirements; Computer Society; Feature extraction; Filters; Handwriting recognition; Image recognition; Image segmentation; Image storage; Noise shaping; Pixel; Smoothing methods;
Journal_Title :
Pattern Analysis and Machine Intelligence, IEEE Transactions on
