Iris Identification Algorithm Research And System Implementation

Iris recognition is an emerging biometric technology, because of the iris ofuniqueness, stability, non-invasive, etc., making it more and more domestic andforeign researchers to focus on. With iris recognition technology, iris recognitionhas made considerable progress in recent years in technology research andapplication development, and has shown broad prospects and great marketdemand. How to design an embedded iris recognition system using digital signalprocessor (DSP) is discussed in the article.First of all, this article has designed a low-cost iris acquisition device. Thedevice consists of a cheap CCD camera, together with the design of near infraredLED ring light. The iris image in the CCD camera is very clear so that thedemand of iris recognition is satisfied. In order to implement non-contactacquisition, image quality evaluation algorithm is also designed.Secondly, the paper introduces a system based on the iris acquisition device,TMS320DM642 EVM and video display device. By video port, decoderconfiguration, we completed development and debugging of hardware and driverson the iris recognition system based on DSP. The entire system can implementvideo input and output functions.After completing the hardware configuration, the author referred to therelated literature, and proposed an improved fast iris recognition algorithm. Theeffective method of locating is to locate pupil firstly and then iris. This method isa combination of rough positioning and precision positioning. The next procedureis iris normalization. At last we extract iris texture using Gabor filter banks, andimplement iris match by relativity function.By DM642 experimental platform, we designed iris recognition systemsoftware based on DSP/BIOS. Software was based on module design, includingsystem initialization, image capture, image processing, and image display module.Subsequent algorithm improvements can easily replace the different imageprocessing algorithm module, but other parts of the software the same.Finally, we made the comparison of this system of experimental results andconclusions of literature. Both rapidity and accuracy are analyzed; advantages and disadvantages of this system are evaluated.