A Generalized iterated shrinkage algorithm for non-convex sparse coding
From zhanglei
Sparse coding with non-convex Lp-norm minimization (0<p<1)
better than Sparse coding with convex L1 norm miminization.
non-convex Lp-norm sparse coding算法有:
1、iteratively reweighted least squares
2、Iteratively thresholding method
3、look-up table
通过扩展soft-thresholding operator, 提出A generalized iterated shrinkage algorithm,解决非凸的稀疏编码问题。