Performance Improvement of Face Recognition System by Decomposition of Local Features using Discrete Wavelet Transforms

As one of the most sought after applications of image analysis, face recognition has received significant attention, especially during the past two decades. Automatic human face recognition has received substantial attention in the recent few years, from researchers in biometrics, pattern recognition and computer vision communities, owing to the high demand for advanced security and authentication needs. Although the existing automated machine recognition systems have certain level of maturity but their accomplishments are limited due to real time challenges. For example, face recognition for the images which are acquired in high contrast with different levels of illumination is a critical problem. Various applications in defense and commercial areas demand real time and high level precision face recognition systems. In turn accuracy involves many floating point operations which will be costly as well as complex in terms of implementation. The major metric in modeling the performance of a face recognition system is its accuracy of recognition. This paper proposes a novel method of face recognition using de-correlation of local features using Discrete Wavelet Transforms (DWT) which improves the recognition accuracy. It also avoids generalizability problem which is caused due to subspace discriminant analysis or statistical learning procedure by using a non-statistical procedure which avoids training step for face samples. This proposed method performs well with images with partial occlusion and images with lighting variations as the local patch of the face is divided into several different patches.