Title :
Matching based on properties of pixel connections for handwritten digit recognition
Author :
Zhu, Xuefang ; Cheng, Minteh
Author_Institution :
Inst. of Acoust., Nanjing Univ., China
Abstract :
We propose a new matching method for differentiating an input pattern from those templates in the definite categories. The method is based on the topological properties of pixel connections. It is called the distance computation method based on tiny lines and their neighbours as comparing regions (DCMBTLNCR). The authors select this modified distance rule (MD) to constitute a classifier
Keywords :
character recognition; edge detection; handwriting recognition; image segmentation; pattern classification; pattern matching; comparing regions; distance computation method; handwritten digit recognition; input pattern; modified distance rule; neighbours; pattern classifier; pattern matching method; pixel connections; templates; tiny lines; topological properties; Acoustics; Character recognition; Hamming distance; Handwriting recognition; High definition video; Information science; Mathematics; Partial response channels; Postal services; Size measurement;
Conference_Titel :
Signal Processing, 1996., 3rd International Conference on
Conference_Location :
Beijing
Print_ISBN :
0-7803-2912-0
DOI :
10.1109/ICSIGP.1996.566555
