DCT Based Information-Preserving Pooling for Deep Neural Networks

Pooling is used in most of deep convolutional neural networks as a feature downsampling method, in order to reduce computation complexity and increase the receptive field size. Since the traditional max/average pooling layers tend to cause severe information loss, spectral pooling with discrete Fourier transform (DFT) is considered as an advisable alternative. In this paper, we propose a novel 2D-discrete cosine transform (2D-DCT) based pooling method for deep neural networks. Due to the energy compaction property, DCT pooling preserves considerably more information than DFT. Moreover, inspired by the separability of DCT, we precompute the transform matrices and embed them into linear layers for parallelization and acceleration on GPUs. Experimental results indicate that the proposed DCT based pooling layer outperforms previous pooling methods on multiple image classification datasets with negligible extra time consumption.