A Content Based Feature Combination Method for Face Recognition

In the last few years, Content Based Image Retrieval (CBIR) system, where images are searched based on their visual contents instead of annotated texts, has drawn enormous attention of researchers because of its growing demand from real world applications. According to many, biometric traits recognition is one of the most potential applications of CBIR. However, very few works have been published on CBIR based face recognition systems. In this research, a content based face recognition process, where color, texture, and shape features are combined to enhance the retrieval accuracy of the system, is proposed. Methodology: 0.4 0.5 0.6 0.7 0.8 0.9 1 Color Histogram [1] Affine Moment Invarinat [3] Gabor Filter [2] Proposed Method A v e r a g e P r e c i s i o n R a t e Grayscale Color Critical Query Fig 3: Comparison of average precision rate of different methods for different databases. Three well known and computationally efficient methods: color histogram [1], Gabor filter [2], and affine moment invariant [3] are used to extract color, texture, and shape features, respectively. Fig. 1 depicts the block diagram of the proposed method. The gray blocks indicate the novel components of the proposed method. Conclusion: A novel content based face recognition method is proposed. Our feature fusion method has better performance than single feature based methods. This method can be applied to any database effectively because of its high recognition rate, ease of computation, and easy weight adjustment features. Future work includes development of a weight learning system for different features.