Design And Research Of Face Recognition Based On NiosⅡ

These years, biology character recognition research have developed fast. In all the technology of biology character recognition, face recognition get more attention because of universality,easy gather and acceptable. More and more scientific research institute and technologist have thrown themselves into the research field of face recognition since last century, and they have get some delighted achievement. The technology of face recognition is using instrument to analyse face image, pick up useful information from it and carry its point of identity identification. Because of broaden application foreground,human face recognition has been the research hotspot in the field of pattern recognition and computer vision.This paper bases on the related research and the latest technique of domestic and international. In this foundation, using the mature image process arithmetic to pre-process for human face image, thereby make the human face location coming true. Then use the 2 Dimension Discrete Cosine Transform to pick up human face character from the image which is locating the human face before. Training human face image swatch by Embedded Hidden Markov Model, and set up the human face database based on EHMM. Let the character matching with the database, then make the face recognition coming true finally.In the field of aptitude instrument, it usually uses DSP as the control core in digital image process. But complicated periphery circuit, long design and debug period and its unsatisfied expansibility are the most shortcoming of the system based on DSP. This paper adopts SOPC which is the latest technology of modern electron design, and designs the human face recognition system based on NiosⅡ. SOPC has integrated the most function module, such as I/O, NiosⅡ, RAM, Flash and user logic into one FPGA, made up a programable system on chip. Using custom instructionsand DSP module which set up by DSP Builder integrated with Matlab in NiosⅡto fulfill human face recognition arithmetic which profit from the function of hardware cooperate with software. It has greatly enhanced the speed of data process and ensured the real time process. The use of free ip core and hardware module in IDE have greatly improved project efficiency and agility.This paper set up the face recognition platform based on DE-2 develop board produced by Terasic Technologies. The experiment shows that the design of face recognition applied SOPC based on NiosⅡhas innate superiorities such as project exploitation cycle, system agility and arithmetic implemented by hardware.this embed system based on NiosⅡhave fulfill the task of face recognition.