Minimum Variance Method to Obtain Best shot in Video for Face Recognition
Kazuo Ohzeki, Ryota Aoyama, Yutaka Hirakawa
Citation: Proceedings of the 2015 Federated Conference on Computer Science and Information Systems, M. Ganzha, L. Maciaszek, M. Paprzycki (eds). ACSIS, Vol. 5, pages 869–874 (2015)
Abstract. Face recognition algorithm using feature points of face parts which is classified in a feature-based methods. As recognition performance depends on the combination of adopted feature points, we utilize all reliable feature points effectively. From input moving video, well-conditioned face images with frontal direction and without facial expression should be extracted. To select such well-conditioned images, an iteratively minimizing variance method is adopted to variable input face images. This iteration drastically brings convergence to the minimum variance 1 for quarter parts of all data, which means 7.5Hz as time scale on average. Also, the maximum interval, which is the worst case, between the two data with minimum deviation is about 0.8 seconds for a tested feature point sample.