The Application And Research On Thenar Palmprint Image Processing

It's a creative work that biometrics combine with Chinese traditional palmprint theory to diagnose disease in humans. Through years of clinical observation and medical research, experts find that thenar palmprints of asthmatics are rough and divide them into four levels based on the rough degree. Using computer technology to do thenar palmprint image processing and quantitative identification can realize the automation of palm old. As an important module of the thenar palmprint quantitative identification and application research of asthma correlation and auxiliary diagnosis, the Thenar palmprint image processing and feature extraction are did in the thesis.In this paper, the palmprint images sampling equipment which based on DSP is developed. The equipment is designed using DSP as its core in order to sample and store the palmprint images. Palmprint image library is established though sampling many palmprint images by the equipment.Get the thenar area orientation using the method of invariant feature points. The method used the valley points between fingers as feature points to look for the locating points and establish coordinate system, then getting the thenar area finally. Finish the preprocessing of thenar palmprint image using wavelet threshold denoising method and the enhancement algorithm based on high frequency emphasize filter and contrast limited adaptive histogram equalization (CLAHE) or wavelet transform. After preprocessing, the texture and grid shape distribution are more obvious. And it's convenient for feature extraction.Using edge extraction methods to extract the line features of thenar palmprint. The number of the line features is more, and its distribution is irregular. Due to the requirements of line features is no unified standard, the number of line features cannot as one of classification standards. A better pattern for the thenar palmprint is that extract the lines features as much as possible, then judge the levels of the thenar palmprint through observing the general number of line features. The binary thenar palmprint thinned image after preprocessing is getted by the serial algorithm based on 4-connectivity parallel thinning algorithm. Extract the node domain feature on the basis of refinement post-processing. The node domain feature as the standard of thenar palmprint classification. By determining the number of node domains to complete the classification of thenar palmprint.