Local Enhancement for Robust Face Detection in Poor SNR Images

Face detection is an important preprocessing stage for automated facial recognition systems. Imaging in real world environment experience substantial variation in lighting and weather conditions. Surveillance systems are a typical example of real world imaging systems. Face detection is based upon typical haar like patterns appearing in facial images. Performance of even most successful face detection algorithms such as viola's work, typically fail in poor lighting or in low SNR (Signal to Noise Ratio) images. Therefore subsequent stage of image enhancement after image acquisition is employed to enhance the faces or face patterns in low lighting images. Proposed research work addressed the face detection in automotive environment. Algorithm addressed the situation of low lighting condition and enhanced the images captured in low lighting conditions with lights switched on. Faces of people in automobile cabin go in complete darkness and difficult to enhance. Proposed scheme is based upon local enhancement method and considerable results are achieved.