![]() | Only 14 pages are availabe for public view |
Abstract Detection of Human faces within images plays an important role in many facial image- related applications such as face recognition/verification, facial expression analysis, pose normalization, and 3D face reconstruction. The performance of these applications is usually to a large degree dependent on accuracy of face detector. Detection of faces is easy for human beings; however, for machines it is not an easy task at all. The difficulty comes from high inter-personal variation (e.g., gender, race), intra-personal changes (e.g., pose, expression), and from acquisition conditions (e.g., lighting, image resolution). Despite the considerable amount of the previous studies on the subject, face detection is not completely solved and it remains a very challenging task. The existing methods need improvements in their accuracy, and a novel robust method that can work under various imaging conditions are required. In this thesis, we propose a new approach that brings us a step closer to this goal. The principal objective of this thesis is to investigate a novel approach toward robust automatic face detector in uncontrolled imaging conditions. We propose an efficient approach that efficiently combines generalized Hough transform by Random decision Forests. Random forests (RF) have become a popular technique for classification, prediction, studying variable importance, variable selection, and outlier detection. There are numerous successful applications examples of RF in a variety of fields, such as object detection, pedestrian detection, and object tracking and action recognition. Random Forests are probabilistically efficient technique that can operate quickly over large datasets. |