Abstract:
In various real-world face recognition applications such as forensics and surveillance, only partial face image is available. Hence, template matching and recognition are strongly needed. In this paper, a genetic algorithm to match a pattern of an image and then recognize this image by this pattern is proposed. This algorithm can use any pattern of an image such as eye, mouth or ear to recognize the image. The proposed genetic algorithm uses a small length chromosome to decrease the search space, and hence the results could be obtained in a short time. Two datasets were used to test the proposed method which are AR Face database and LFW database of face, the overall matching and recognition accuracy were calculated based on conducting sequences of experiments on random sub-datasets, where the overall matching and recognition accuracy was 91.7% and 90% respectively. The results of the proposed algorithm demonstrate the robustness and efficiency compared with other state-of-the-art algorithms.
