The algorithm studies of Hidden Markov Model in face distinguishing

Hidden Markov Models(HMM) have been successfully used in speech recognition where data is essentially one-dimensional. An new approach is proposed. In this approach, Dauechies orthogonal wavelet transform is used to preprocess the original face image, resulting in its four sub-images belonging to different frequency bands, and the sub-images are used to learning and recognition based on HMM. A algorithm is designed to combine the multiple sort results, and the Karhunen Loeve Transform (KLT) was used to extract a set of observations that improving the method by Asmaria. This approach increases the ratio of recognition and reduces the time of computing. The experimentations prove the approach is rational.