ON ROBUST AUTOMATIC OFFLINE ISOLATED HANDWRJTTEN WORD VERIFICATION ALGORITHM

Automatic Offiine Isolated handwriting word verification system using template matching (pixel by pixel matching) is proposed in the research work. The composition of 400 word essay, the handwritten biometrics of 50 people (students); 25 males and 25 females to handwrite the essay five times in separate sheet of papers the aim here is to verify if a particular handwriting is from that same person or is a fake. No speclfzcation for pen type and color were given. The collected documents were digitized and stored in a .bmp files. Different pre-processing stages such binarization, noise removal, thinning, line and word segmentation were carried on the file before matching the files. The use of template matching considering pixel by pixel matching to do exhaustive test on different files to get a match. The results were quite impressive different channels distortion were gotten based on the similarity and dissimilarity tests. For the similarity test all channel distortion are Zero (0) meaning that the files are the same, while in the dissimilarity test channel distortion varies telling us that the files are ji-om d[fferent people.