A finely trimmed input datasets to svm is formed to get a higher recognition rate and speed . the combination of svm and ica is highly effective in the recognition process . the dimension reduction of ica is the fundamental factor in the recognition rate and speed improvement and the modified svm method gives an amazing result in this area Ica方法的引入对于识别过程无论在速度上还是在正确率上都有很大的提高，其中的根本原因在于对输入数据维数的压缩；修正svm方法是对原有svm方法的一个改进，借助于特征压缩后的数据集，识别效率有了很大的改善。
