Application of neural networks in name card classification
Sheung-Lun Hung, Ken C. K. Law, Lam For Kwok · 2002
Name cards generally consists of figures, Chinese and English characters. The English characters can be in different fonts and various sizes. A name card storage and retrieval system which can discriminate the segments of interest on the name card and using neural networks for character recognition has been developed. A semantic field recognition technique has also been implemented to allow the automatic storage of name card information in a popular database system. The segmentation technique used in the discrimination process is found to be very satisfactory. The system is extremely efficient in terms of space and speed.>