Adaptive Context Network for Scene Parsing

TL;DR

This paper introduces ACNet, an adaptive scene parsing network that dynamically fuses global and local contexts per pixel, leading to state-of-the-art results across multiple datasets.

Contribution

It proposes a novel pixel-aware context fusion method using adaptive modules that measure and combine global and local context demands for improved scene parsing.

Findings

Achieves new state-of-the-art performance on Cityscapes, ADE20K, PASCAL Context, and COCO Stuff datasets.

Demonstrates the effectiveness of adaptive context fusion over fixed context methods.

Validates the approach through comprehensive experimental evaluations.

Abstract

Recent works attempt to improve scene parsing performance by exploring different levels of contexts, and typically train a well-designed convolutional network to exploit useful contexts across all pixels equally. However, in this paper, we find that the context demands are varying from different pixels or regions in each image. Based on this observation, we propose an Adaptive Context Network (ACNet) to capture the pixel-aware contexts by a competitive fusion of global context and local context according to different per-pixel demands. Specifically, when given a pixel, the global context demand is measured by the similarity between the global feature and its local feature, whose reverse value can be used to measure the local context demand. We model the two demand measurements by the proposed global context module and local context module, respectively, to generate adaptive contextual features. Furthermore, we import multiple such modules to build several adaptive context blocks in different levels of network to obtain a coarse-to-fine result. Finally, comprehensive experimental evaluations demonstrate the effectiveness of the proposed ACNet, and new state-of-the-arts performances are achieved on all four public datasets, i.e. Cityscapes, ADE20K, PASCAL Context, and COCO Stuff.