clcNet: Improving the Efficiency of Convolutional Neural Network Using Channel Local Convolutions

Depthwise convolution and grouped convolution has been successfully applied to improve the efficiency of convolutional neural network (CNN). We suggest that these models can be considered as special cases of a generalized convolution operation, named channel local convolution(CLC), where an output channel is computed using a subset of the input channels. This definition entails computation dependency relations between input and output channels, which can be represented by a channel dependency graph(CDG). By modifying the CDG of grouped convolution, a new CLC kernel named interlaced grouped convolution (IGC) is created. Stacking IGC and GC kernels results in a convolution block (named CLC Block) for approximating regular convolution. By resorting to the CDG as an analysis tool, we derive the rule for setting the meta-parameters of IGC and GC and the framework for minimizing the computational cost. A new CNN model named clcNet is then constructed using CLC blocks, which shows significantly higher computational efficiency and fewer parameters compared to state-of-the-art networks, when being tested using the ImageNet-1K dataset.