Title :
Multi-experts for touching digit string recognition
Author :
Wang, Xian ; Govindaraju, Venu ; Srihari, Sargur
Author_Institution :
Center of Excellence for Document Analysis & Recognition, State Univ. of New York, Buffalo, NY, USA
Abstract :
84.6% of touching digit strings have only two digits touching, 12.3% have three digits touching and 3.1% have more than three digits touching. We present a multi-expert approach to recognize touching digit pairs (TDP) and touching digit triples (TDT). We combine holistic and traditional segmentation methods. 25,686 TDP training samples and 2,778 TDP testing samples collected from USPS mail are used in our experiment. The holistic method outperforms the traditional segmentation-based methods. The multi-expert combination has the best performance: a correct recognition rate of 91.1% on TDP
Keywords :
expert systems; image segmentation; multi-agent systems; optical character recognition; postal services; software performance evaluation; US Postal Service; USPS mail; character segmentation methods; holistic method; multi-expert approach; performance; touching digit pairs; touching digit string recognition; touching digit triples; training samples; Histograms; Image analysis; Image segmentation; Labeling; Postal services; Read only memory; Testing; Text analysis; Venus;
Conference_Titel :
Document Analysis and Recognition, 1999. ICDAR '99. Proceedings of the Fifth International Conference on
Conference_Location :
Bangalore
Print_ISBN :
0-7695-0318-7
DOI :
10.1109/ICDAR.1999.791909