Location Algorithm Based On Rough Sets And Gray Level Histogram Of The Human Eye

The technology of face recognition is one of the hottest research directions for its wide application and urgent demand. A general face recognition system mainly consists of five parts, named pretreatment, face location, feature extraction, sample study and sample recognition process. Among them, the most important steps, i.e., face location and feature extraction will directly affect the effect of recognition. Most face location methods are based on the complexion model. And the eye location in the feature extraction has a very significant reference effect.Therefore, aiming at the character that eyes have relatively low gray degree compared to whole face, we propose a new algorithm of eye location based on Rough Sets and gray-value histogram, which is on the basis of face rough location, and highly promotes the validity of eye location because of narrowing the search range.In this paper, some most popular methods of pretreatment to face images are briefly introduced. Some typical and effective methods of feature extraction used in face recognition study are discussed, such as methods of geometry feature extraction, methods of algebra feature extraction, methods based on neural network and so on. In Section part , we discuss the typical face dectecting method based on the complexion model and In three part, some algorithms of eye location in common use are mentioned, including Hough Transform, Deformation template and Edge feature Analysis. At last, we propose an algorithm of eye location based on rough set and gray-value histogram, and gives detailed experimental data.The experiment results proves that the algorithm of eye location based on rough set and gray-value histogram has 90 percent accuracy rate of eye location in the existing experimental data. Moreover, with the increase of object attribute data, the accuracy rate of eye location can be further improved.