Real Time Face Recognition using Steerable Filters and Template Matching

Main Article Content

Sunil Kumar
Sakshi Gupta


In this paper a technique of Face Recognition is discussed. Face detection in the image is an important research branch of face recognition. For the purpose of recognising the faces in images efficiently, a face recognition method based on face-template is proposed. According to the character of density of the feature organs such as eye, ear, nose, mouth, part of cheek in the face images, the frontal full face-template is constructed. And the frontal half face template is constructed directly based on density symmetry of face-template. Template formation required to figure out the region containing facial components. This is done using steerable filters. Eyes usually contain both dark and bright pixels in the luma component. Gray scale morphological operations like dilation and erosion are performed on the image to emphasize brighter and darker pixels in the luma component around eye regions. We further notice that the mouth has a relatively lower response in the Cr/Cb feature but a high response in Cr2. Therefore, the difference between Cr2and Cr/Cb can emphasize the mouth regions. The region containing the eye and mouth region is taken as templates. The face detection experiments in the frame images from the surveillance video were carried on using the method of colour segmentation and face is recognised according to comparability between template and detected face. The theory analysis shows that the average face-template can reduce the chanciness of local density of the template effectively and the half face-template can reduce the symmetry redundancy of density in the face-template and increase the speed of face detection. The experimental result indicates that the half face-template can adapt to side face images in a large angle, which improves the correctness of side face recognition substantially. 

Keywords: Template matching, Face Detection, Face Recognition, Steerable Filters, Comparability.



Download data is not yet available.

Article Details