MASSIVELY-PARALLEL HANDWRITTEN CHARACTER-RECOGNITION BASED ON THE DISTANCE TRANSFORM

A new statistical classifier for handwritten character recognition is presented. After a standard preprocessing phase for image binarization and normalization, a distance transform is applied to the normalized image, converting a black and white (B/W) into a gray scale picture. T he latter is used as feature space for a k-Nearest-Neighbor classifier , based on a dissimilarity measure which generalizes the use of the di stance transform itself. The classifier has been implemented on a mass ively-parallel processor, Connection Machine CM-2. Classification resu lts of digits extracted from the U.S. Post Office ZIP code database an d the upper-case letters of the NIST Test Data 1 are provided. The sys tem has an accuracy of 96.73% on the digits and 94.51% on the upper-ca se letters when no rejection is allowed and an accuracy of 98.96% on t he digits and 98.72% on the upper-case letters at 1% error rate.