Research Of Iris Recognition Algorithm Based On Wavelet Transform And Its Realization Of DSP

Biometric recognition technology is used more and more widely in modern society, iris feature is inviolability, uniqueness and more stable in comparison with other human organ feature(face, finger print,gait and so on)with further research on this area, the potential value of iris recognition is being exploited and more and more scientific researchers begin to focus on it.The whole iris recognition process includes image acqusistion, iris image preprocessing, feature extraction and matching. This paper puts a lot of emphasis on the phase of feature extraction and matching, and realizes the iris feature extraction based on 1-D discrete wavelet transform and based on 2-D discrete wavelet transform. and then to classify iris with weighted Hamming distance. In the end this paper realizes the whole iris recognition algorithm, and makes it run under the code composer studio platform.In the aspect of the most important iris recognition algorithm, this thesis takes stress on researching the wavelet transform theory and the way of extract iris feature with wavelet tool.With matlab tool,this paper makes decision of which wavelet base and wavelet decomposition level is better. And give the effective way of how to denoise the iris and how to solve the rotational inconsistencies. In the aspect of realization the classifier, this paper takes advantage of weighted hamming distance (WHD) instead of minium hamming distance(MHD).In the aspect of the DSP realization of the iris recognition, this thesis makes analysis of the difference between matlab grammar and C grammar,and point out the environment difference betweeen the PC and DSP when running the program. Then realizes the iris recognition algorithm under the DSP hardware.In the contents of code optimization, this thesis introduces optimization strategies and methods of the C language program under DM6437, and gives the detail execution steps of loop optimization with an example. Finally, this paper gives optimization result and the comparision of part of algorithm before optimization and after optimization by the consuming time.