Improved Efficient Dictionary Learning with Cross-Label and Group Regularization
Existing works have revealed that dictionary learning can effectively preserve the label property when applied to signal reconstruction and face recognition. As time goes, the methods based on dictionary learning become increasingly popular due to their superior accuracy and efficiency. Based on this, an improved dictionary learning model is proposed in this paper to find the balance between the time cost of operating the algorithms and the residuals generated when reconstructing signals with the learnt dictionary sparse codes. Demonstrated by the results of experiments, the proposed method intends to determine a more reasonable sparse dimension, which can not only obtain desired classification results but also decrease much of the redundancy in experimental computation.
KeywordsCross-label suppression Computational complexity Dictionary learning Face recognition Compressive sensing
- 4.Wang, F., Lee, N., Sun, J., Hu, J., Ebadollahi, S.: Automatic group sparse coding. In: Proceedings of the Twenty-Fifth AAAI Conference on Artificial Intelligence (AAAI). San Francisco, California, USA, 7–11 August 2011Google Scholar
- 8.Zhang, Q., Li, B.: Discriminative K-SVD for dictionary learning in face recognition. In: Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, pp. 2691–2698 (2010)Google Scholar
- 9.Ramirez, I., Sprechmann, P., Sapiro, G.: Classification and clustering via dictionary learning with structured incoherence and shared features. In: Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, pp. 3501–3508 (2010)Google Scholar