KCNN: Kernel-wise Quantization to Remarkably Decrease Multiplications in Convolutional Neural Network.
Linghua ZengZhangcheng WangXinmei TianPublished in: IJCAI (2019)
Keyphrases
- convolutional neural network
- face detection
- pairwise
- support vector
- feature space
- kernel function
- kernel regression
- input space
- mutual subspace method
- sparse kernel
- kernel density estimation
- quantization error
- kernel pca
- component analysis
- neural network
- similarity function
- kernel methods
- feature extraction
- machine learning
- kernel learning
- string kernels
- image compression
- kernel space
- uniform quantization
- computer vision