Automatic 3d Face Recognition From 2d Image through Projection

Abstract

As one of the most successful applications of image analysis and newlineunderstanding, face recognition has recently received significant attention, newlineespecially during the past several years. Although face recognition is not as newlineaccurate as the other recognition methods, it still grabs huge attention of many newlineresearchers in the field of computer vision. The main reason behind this newlineattention is the fact that the face is the conventional way people use to identify newlineeach other s and without user cooperation we can recognize them. newlineObjective of this research work is two folds: newline(1) A technique for creating a 3D model of face its front and side views newlineand, newline(2) A method for 3D based recognition of faces. newlineTo fulfill our first objective software is developed. In this work we created a newlinenew parameterized face model based on facial features i.e. front and side view newlineof face, which make it enables for applicant to estimate the correct parameters newlinefor a specific human face. This model defines vertex weights for each and newlineevery front and side facial feature, which gives us the information about newlinetransformations of each vertex. We used a method for detecting facial newlineparameters form images with Haar Cascade and Skin Color based image newlinesegmentation in this work. The method gives us correct results to achieve newlinegood precision with updating a face model. We give a comparative study of newline3D face reconstruction methods and show that our method is computationally newlineefficient. We than propose a face recognition scheme on the reconstructed newlineimage. newlineTo fulfill our second objective for this purpose we first detect the newlinehuman faces from a still image using face detection algorithm and next to newlinerecognize the detected faces from the training database. The algorithms used newlinefor these purpose are: Ada boost with Haar classifier face detection algorithm newlineas the name suggests for face detection purpose and Eigenfaces algorithm for newlinethe face recognition purpose. We have experimented with web-cam images as newlineinput and detected as well as recognize faces in it.