Exploiting Kernel Sparsity and Entropy for Interpretable CNN Compression.
Yuchao LiShaohui LinBaochang ZhangJianzhuang LiuDavid S. DoermannYongjian WuFeiyue HuangRongrong JiPublished in: CVPR (2019)
Keyphrases
- cellular neural networks
- image compression
- data compression
- information theoretic
- compression ratio
- information theory
- kernel methods
- mutual information
- compression algorithm
- feature space
- high dimensional
- kernel function
- sparse representation
- compression scheme
- convolutional neural network
- support vector
- reproducing kernel hilbert space
- linear combination
- hidden markov models
- lossless compression
- multi class
- information entropy
- compression rate
- positive definite
- kernel regression
- shannon entropy
- image processing
- adaptive binary arithmetic coding