Learnable Polyphase Sampling for Shift Invariant and Equivariant Convolutional NetworksDownload PDF

Published: 31 Oct 2022, Last Modified: 27 Dec 2022NeurIPS 2022 AcceptReaders: Everyone
Keywords: Convolutional Neural Networks, Shift equivariance, Shift invariance, Polyphase Decomposition
Abstract: We propose learnable polyphase sampling (LPS), a pair of learnable down/upsampling layers that enable truly shift-invariant and equivariant convolutional networks. LPS can be trained end-to-end from data and generalizes existing handcrafted downsampling layers. It is widely applicable as it can be integrated into any convolutional network by replacing down/upsampling layers. We evaluate LPS on image classification and semantic segmentation. Experiments show that LPS is on-par with or outperforms existing methods in both performance and shift consistency. For the first time, we achieve true shift-equivariance on semantic segmentation (PASCAL VOC), i.e., 100% shift consistency, outperforming baselines by an absolute 3.3%.
Supplementary Material: zip
12 Replies