KODAK lMAGELINK™ OCR Alphanumeric Handprint Module
Alexander Shustorovich, Christopher W. Thrasher · Neural Information Processing Systems · 1995
This paper describes the Kodak Imagelink™ OCR alphanumeric handprint module. There are two neural network algorithms at its core: the first network is trained to find individual characters in an alphanumeric field, while the second one performs the classification. Both networks were trained on Gabor projections of the original pixel images, which resulted in higher recognition rates and greater noise immunity. Compared to its purely numeric counterpart (Shusutorovich and Thrasher, 1995), this version of the system has a significant application specific postprocessing module. The system has been implemented in specialized parallel hardware, which allows it to run at 80 char/sec/board. It has been installed at the Driver and Vehicle Licensing Agency (DVLA) in the United Kingdom, and its overall success rate exceeds 96% (character level without rejects). which translates into 85% field rate. If approximately 20% of the fields are rejected, the system achieves 99.8% character and 99.5% field success rate.