Publication Date


Degree Program

Department of Mathematics and Computer Science

Degree Type

Master of Science


The process of detecting a face from a video in real time is essential in applications such as human surveillance, human computer-interaction, and for further face recognition research purposes. In this paper, the face detection algorithm is divided into four stages namely, Video Database Acquisition (VDA), Frame Sequence Extraction (FSE), Skin Region Detection (SRD), and K-Mean Face Segmentation (KFS). Initially, the videos in MPEG format are converted to JPEG images depending on the user specified frame rate (FSE phase). During this conversion, the face detection process comprising of SRD and KFS phases runs on each of the images that are converted. The skin regions are detected in the images, which act as the input for the K-Mean Face Segmentation phase. The skin region clusters thus obtained are classified as face clusters depending on a threshold value. This algorithm was tested on 18 videos, which were acquired by the SONY DCR TRV-80 camera in the VDA phase, regardless of age, gender, size, race, and skin tones. Furthermore, the varying illumination conditions such as bright sunlight, sufficient light, and dim light conditions, and different orientations of the individuals in the videos were gracefully handled by the system. The time taken to detect and store the normalized faces was comparable to the length of the video and in some cases it was even less. Thus, this system works in True Real Time (TRT).


Computer Sciences