An Automated Mail Sorter System using SVM Classifier
K.S. Arun, Jerin Thomas · International Journal of Computer Applications · 2011
paper proposes an Automated Mail Sorter (AMS) system that scans a mail and interprets one of the imperative fields of the destination address, the pin code to sort the mails. The scanned document was segmented into different fields to extract the pin code. A general classifier for the recognition of pin-code digits written in English was then employed. The recognition system consists of a feature extractor and a classification network. The feature extracted was Hu moments and the classification network was the support vector machine using a polynomial kernel of order 2. Other classification networks such as the multi feature recognizer and decomposing network were also used, but SVM gave the maximum accuracy. General Terms