The Research Of Image Segmentation Based On Support Vector Machine

With the development of social and economy,the accuracy,security and utility of personal identification methods are put forward a highly requirement. The traditional personal identification methods based on token or password have the hidden troubles. The token or password is prone to be lost, forgotten,copied and stolen. The traditional personal identification methods cannot satisfy the security requirement of our highly interconnected information society. Biometrics based on the physiological or behavior traits identification provides a convenient and reliable scheme. Among the identification technologies, the technique of fingerprint identification has become one of the widest used biometric identification techniques due to its convenience, high accuracy and low cost. It has been widely used in electronic commerce, criminal identification, information safety etc. Image segmentation is an important part of the image pre-processing, effective segmentation not only can reduce the follow-up processing time but also enhance the reliability of the feature extraction.In this paper, we analyze previous algorithms and present a fingerprint image segmentation algorithm based on support vector machine(SVM). At first, the image is partitioned into blocks and the low gray variance background blocks were segmented by the contrast. After that, the gray mean, the gray variance, the contrast, the coherence and the main energy ratio were extracted on the remaining blocks. We use supervised SVM to classify patterns and select typical patterns to train the classifier. The blocks that can not be decided by the first segmentation were segmented by a SVM classifier. Finally, morphology was applied as postprocessing to reduce the number of classification errors. The simulations were done to the proposed algorithms on FVC database, and the experiments show that the proposed method is effective and robust.