A Two-step Feature Extraction Algorithm for Face Recognition
Xuansheng Wang, Weibin Zeng, Hongying Zheng, Tangren Dan, Huazhong Li, Jianqiang Sheng · 2020
Feature extraction is very important for face recognition. In the past, most of the methods directly transform samples from a high-dimensional space to a new low-dimensional subspace. In this paper, we propose a novel method for feature extraction. This method can get a new sample from the original sample, and the size of the new sample is much smaller than the original sample, but keep the core information of the original sample. Since this method can be combined with any feature extraction method after this step; we call it two-step feature extraction method. From experiments, we can find that our method is effective and efficient.