Face detection

Abstract

Human face detection plays an important role in applications like face recognition, video surveillance, human computer interface, face image database management and many more. In modern multimedia systems, video and image signals usually need to be indexed or retrieved according to their contents. In our thesis, we implement a color characteristic for use in detection of frontal human faces in color images with complex backgrounds i.e. a color based technique to detect frontal human face had been developed and implemented. A technique for detecting frontal human faces in color images is described that first separates skin region from non-skin region and then locates faces within skin regions. Using color information in an image is one of the various possible techniques for face detection. The technique involves conversion of a color image into a gray scale image in such a way that the gray values in the pixel shows the likelihood of the pixel belonging to the skin. Obtained gray scale image is then segmented to skin and non-skin regions, and a model face, representing front face is used in template matching process to detect face within skin regions i.e. to find which of the candidates is/are actually a face. Later, the false-positive and false-negative errors of the implemented face detection technique on color images are calculated. The experimental results show that this method can detect faces in v the images from different sources with good efficiency. Since faces are common elements in video and image signals, the proposed face detection technique is an advance towards the goal of content-based video and image indexing and retrieval.