Detecting faces across multiple views is more challenging than in a fixed view, e.g. frontal view, owing to the significant non-linear variation caused by rotation in depth, self-occlusion, and self-shadowing. The view sphere is separated into several small segments. On each segment, a face detector is constructed. Estimating the pose of an image regardless of whether it is a face a pose estimator is constructed using Support Vector Regression. PCA is adopted to represent multi-view face patterns in a low-dimensional orthogonal feature space, and SVM regression is employed to construct the pose estimators. The advantage of the SVM pose estimator is that it can be trained directly from the data with the little requirement of the prior knowledge about the data and it is guaranteed to converge. The pose information is used to choose the appropriate face detector to determine if it is a face. With this pose-estimation based method, considerable computational efficiency is achieved. Meanwhile, the detection accuracy is also improved since each detector is constructed on a small range of views. The algorithm to find face detection by combining the Eigenface and SVM methods which performs almost as fast as the Eigenface method but with a significantly improved speed. By combining the two methods together, a novel method is proposed which keeps the advantages and suppresses the disadvantages of both methods. Matlab(2015b) simulation results are shown in the demo video presented including tuning the parameters of the pose estimators, performance evaluation, and applications to face detection and frontal-view face recognition.

Reference Paper: Support Vector Machine Based Multi-View Face Detection and Recognition

Author’s Name: Yongmin Lia,Shaogang Gongb, Jamie Sherrahc,and Heather Liddellb

Source: Elsevier-Image and Vision Computing

Year: 2004

Request source code for academic purpose, fill REQUEST FORM or contact +91 7904568456 by WhatsApp or sales@verilogcourseteam.com, fee applicable.

SIMULATION VIDEO DEMO