Feature Coding Method Based on Shared Weights Support Vector Data Description for Face Recognition

Mengmeng Liao,Yunjie Li,Meiguo Gao
DOI: https://doi.org/10.1088/1742-6596/1955/1/012029
2021-01-01
Journal of Physics Conference Series
Abstract:In this paper, we propose a feature coding method based on shared weights support vector data description (FCM-SWSVDD). The proposed process of FCM-SWSVDD is as follows. By considering the density information of the clusters and introducing the weighting learning, we propose an improved support vector data description (SVDD), named shared weights support vector data description (SWSVDD). SWSVDD can obtain the cluster center and cluster radius more accurately. Incorporating SWSVDD and the triangle coding into the same feature coding learning process, FCM-SWSVDD is proposed. After the features of those images are extracted by using FCM-SWSVDD, a sparse representation classifier is used to classify those features. Experimental results show that the performance of the proposed method exceeds many methods.
What problem does this paper attempt to address?