Extraction Of The Facial Edge And Corner Detection

In the thesis, our purpose is to solve two problems which are used during the course of the visa application in the embassy, the problems are: how to locate a faceand how to detect the eyes' corners. After summarizing the main techniques of person identification and face recognition, we describe and debate the main methods of image smoothing and filtering how to locate the face from a registration imageand how to detect the eyes' corners.The thesis is organized as following : Image smoothing , image filter and edge detection. This is the first partof the thesis. We summarize the main techniques of smoothing and filtering, and describe the methods of edge detection which uses Gaussian function as smoothing function, and an image enhancement and edge detection method which is based on the techniques of image filter , image enhancement and edge detection . Initializing the face's edge and obtaining the eyes' image. In this part. we describe how to calculate the ellipses connecting the edges of head top and head bottom and how to locate the axes connecting the eyes' pupils , brows nose and the face's side limits, after having calculated these parameters, we initialize the face's edge and extracting the eyes' image by using these parameters . Solving the face's accurate edge. This is the central part of the thesis. We first describe "the snake model" and "energy-minimizing function", the latter includes three forces : internal forces , external forces and image forces ; Then ,we debate and change the three forces of energy-minimizing function ; Finally, weuse the greedy algorithm to solve the minimization of energy-minimizing function and decide the face's accurate edge, the greedy algorithm runs much faster and needs less memory than the methods of other authors .