CVAILGDec 21, 2021

RepMLPNet: Hierarchical Vision MLP with Re-parameterized Locality

arXiv:2112.11081v285 citationsHas Code
Originality Highly original
AI Analysis

This work addresses the limitation of MLPs in vision tasks by enabling local pattern capture, making them more competitive with convolutional networks for image recognition and segmentation.

The paper tackles the problem of fully-connected layers being poor at capturing local patterns in image recognition by proposing Locality Injection, a method to incorporate local priors into FC layers via merging trained conv kernel parameters, resulting in RepMLPNet which shows favorable accuracy-efficiency trade-offs and is the first MLP to seamlessly transfer to Cityscapes semantic segmentation.

Compared to convolutional layers, fully-connected (FC) layers are better at modeling the long-range dependencies but worse at capturing the local patterns, hence usually less favored for image recognition. In this paper, we propose a methodology, Locality Injection, to incorporate local priors into an FC layer via merging the trained parameters of a parallel conv kernel into the FC kernel. Locality Injection can be viewed as a novel Structural Re-parameterization method since it equivalently converts the structures via transforming the parameters. Based on that, we propose a multi-layer-perceptron (MLP) block named RepMLP Block, which uses three FC layers to extract features, and a novel architecture named RepMLPNet. The hierarchical design distinguishes RepMLPNet from the other concurrently proposed vision MLPs. As it produces feature maps of different levels, it qualifies as a backbone model for downstream tasks like semantic segmentation. Our results reveal that 1) Locality Injection is a general methodology for MLP models; 2) RepMLPNet has favorable accuracy-efficiency trade-off compared to the other MLPs; 3) RepMLPNet is the first MLP that seamlessly transfer to Cityscapes semantic segmentation. The code and models are available at https://github.com/DingXiaoH/RepMLP.

Code Implementations4 repos
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes