L1-(2D)2PCANet: A Deep Learning Network for Face Recognition
This is an incremental improvement for face recognition systems, enhancing robustness to outliers in training images.
The authors tackled face recognition by proposing L1-(2D)2PCANet, a deep learning network that uses L1-norm-based two-directional two-dimensional principal component analysis to learn convolution filters, resulting in better recognition performance than baseline networks like CNN and PCANet on datasets such as YALE and LFW-a, especially with outliers.
In this paper, we propose a novel deep learning network L1-(2D)2PCANet for face recognition, which is based on L1-norm-based two-directional two-dimensional principal component analysis (L1-(2D)2PCA). In our network, the role of L1-(2D)2PCA is to learn the filters of multiple convolution layers. After the convolution layers, we deploy binary hashing and block-wise histogram for pooling. We test our network on some benchmark facial datasets YALE, AR, Extended Yale B, LFW-a and FERET with CNN, PCANet, 2DPCANet and L1-PCANet as comparison. The results show that the recognition performance of L1-(2D)2PCANet in all tests is better than baseline networks, especially when there are outliers in the test data. Owing to the L1-norm, L1-2D2PCANet is robust to outliers and changes of the training images.