Noisy english character recognition by combining SVM classifier

The support vector machine (SVM) is a new learning machine with very good generalization ability. The SVM classifier has superior recognition rates when compared to other classifiers. The input data is the images of noisy characters. By combining the SVM classifier the recognition of noisy characters is better. This approach works in 2 steps, 1.) feature extraction module, 2.) SVM classifier module. The recognition of SVM with Gaussian kernel with statistical feature is 97.7% and with structural and statistical feature is 99.3%.