Fingerprint Identification Algorithms Based On DSP

Because of uniqueness and invariability, the fingerprint identification is becoming one of the most popular personal authentication technologies, and many fingerprint products have been offered nowadays. Along with the broadness of the fingerprint identification used in the folk field, the research of embedded fingerprint identification is deepened more and more.The two core questions of embedded fingerprint identification are independent system and high speed. TMS320C6416DSP is suitable for parallel algorithm to increase the speed due to its characteristics and can be used as a independent system. This paper chooses this DSP as the embedded system , makes some researches on the fingerprint identification algorithms based on the DSP and proposes some algorithms suitable.(1) In the preprocessing of the fingerprint image ,a parallel fingerprint segment algorithm is proposed according to the characteristics of the fingerprint segment and TMS320C6416DSP ; When calculating direction image, because of many gradient operations, a parallel gradient algorithm based on TMS320C6416DSP is given and due to the characteristic of the directional image an angle calculating algorithm in blocks and with comparison is also proposed to rise the speed; With binarization, this paper puts forward to an improved dynamic threshold binarization algorithm by coding to find the pixels; In the process of thinning, a parallel thinning algorithm based on TMS320C6416DSP is proposed.(2) In the aspect of feature extraction, a fast and effective feature extraction and post-processing algorithm is selected and improved to adapt to TMS320C6416DSP.(3) In the matching section, a matching algorithm based on index triangle is brought forward. Firstly the index triangles are made using the extracted feature and a fuzzy distinguish is following according to the indexing result. When the result is in a certain scope, the matching vector is made up based on the first or second classic triangle to do the local feature matching.All algorithms mentioned in the paper have simulatively been experimented on TMS320C6000CCS2.0 . Results show that these new algorithms given in this paper can significantly improve the computation efficiency with the same results as the original algorithms.