|
Computer face recognition in recent years, a very active area of ??research. It a wide range of applications, such as authentication security system, video surveillance, target recognition and tracking, as well as expression analysis, analysis of age, such as lip reading. Face detection is the basis of face recognition is a key link in automatic face recognition system, accurate and efficient face detection algorithm has a very important role to improve the face recognition rate. This paper studies color face images of the human face detection, face detection needs to face image denoising, edge detection and face image segmentation and image illumination effects processing, to eventually determine a human faces, face image denoising problem, the fast particle swarm optimization algorithm is applied to the wavelet threshold shrinkage threshold denoising method optimization, the image as a particle, the particle updates itself by tracking the two extremes, seek the optimal threshold value, the optimal threshold denoising purpose is to minimize the noise impact of the image, the experimental results prove that this method not only peak noise ratio significantly improved image quality visual has also been improved, and the noise variance The greater the peak noise ratio and image quality, the more it can show its superiority. For edge detection, the paper proposed an improved particle swarm optimization algorithm, and then introduced the quaternion representation of the color image, the face edge detection using this new algorithm, to overcome the defects of the traditional color image edge detection loss problem, solve the edges, so that the effect of edge extraction significantly improve the flexibility and adaptability, the experimental results show that the proposed method has a better effect on the color image edge detection, and to be able to extract many of the traditional color image edge detection method can not be extracted from the image texture detail, and the algorithm is stable, easy convergence faster edge detection speed. For the problem of image segmentation of the human face, an improved algorithm, the image is divided into n × n window, respectively, using two-dimensional Otsu method based on the bee colony algorithm for image segmentation within each window, the purpose of accurate, rapid to find the optimal threshold for image segmentation and image optimal split, the simulation results show that the proposed method can get ideal segmentation results, and the computation is greatly reduced, to achieve the purpose of rapid segmentation to facilitate the two-dimensional Otsu real-time application of the method, which proved that the method is feasible and effective.
|