http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
( Xingjian Gu ),( Xiangbo Shu ),( Shougang Ren ),( Huanliang Xu ) 한국인터넷정보학회 2018 KSII Transactions on Internet and Information Syst Vol.12 No.7
Slow Feature Discriminant Analysis (SFDA) is a supervised feature extraction method inspired by biological mechanism. In this paper, a novel method called Two Dimensional Slow Feature Discriminant Analysis via L<sub>2,1</sub> norm minimization (2DSFDA-L<sub>2,1</sub>) is proposed. 2DSFDA-L<sub>2,1</sub> integrates L<sub>2,1</sub> norm regularization and 2D statically uncorrelated constraint to extract discriminant feature. First, L<sub>2,1</sub> norm regularization can promote the projection matrix row-sparsity, which makes the feature selection and subspace learning simultaneously. Second, uncorrelated features of minimum redundancy are effective for classification. We define 2D statistically uncorrelated model that each row (or column) are independent. Third, we provide a feasible solution by transforming the proposed L<sub>2,1</sub> nonlinear model into a linear regression type. Additionally, 2DSFDA-L<sub>2,1</sub> is extended to a bilateral projection version called BSFDA-L<sub>2,1</sub>. The advantage of BSFDA-L<sub>2,1</sub> is that an image can be represented with much less coefficients. Experimental results on three face databases demonstrate that the proposed 2DSFDA-L<sub>2,1</sub>/BSFDA-L<sub>2,1</sub> can obtain competitive performance.