Facial Boundary Image Reconstruction Using Elliptical Approximation and Convex Hull

摘要

The Stretch of Biometrics' applications can only be limited by limiting ones' imagination! Biometrics is the Science and Technology of measuring and analyzing biological data for authentication purposes. In addition to verification the guiding force behind biometric verification has been convinience. Face enjoys a prime position in the realm of biometrics because it is very easily accessible when compared to the other biometrics. Efficient accomplishment of Face Recognition confronts innumerable hurdles in the form of variations in lighting conditions during image capture, Occlusions, damage in facial portions due to accidents etc. The application of Facial Image Inpainting also fails when the occlusions or the deformalities are present across the boundary of the object of interest(face), since the bounds for the application of the inpainting algorithm is not precisely defined. Hence recovery of the complete picture of a human face from partially occluded images is quite a challenge in Image Processing. The proposed FIREACH algorithm concentrates on the generation of a convex hull and a non linear elliptical approximation of the depleted and partially visible boundary of the human face, given different parameters to achieve an Efficient Boundary Recovery. The Boundary Recovery Algorithm is a pre-processing step which aids in setting up of a suitable platform for the proficient application of the Facial Image Inpainting.