Learnable Tree Filter for Structure-preserving Feature Transform

TL;DR

This paper introduces a learnable tree filter that models long-range dependencies in semantic segmentation while preserving object details, using a novel efficient algorithm for resource reduction and integration into neural networks.

Contribution

The paper proposes a generic, structure-preserving tree filtering module with a linear-time algorithm, improving segmentation performance with less computational overhead.

Findings

Achieves better performance than PSP and Non-local methods

Reduces resource consumption significantly

Provides consistent improvements across benchmarks

Abstract

Learning discriminative global features plays a vital role in semantic segmentation. And most of the existing methods adopt stacks of local convolutions or non-local blocks to capture long-range context. However, due to the absence of spatial structure preservation, these operators ignore the object details when enlarging receptive fields. In this paper, we propose the learnable tree filter to form a generic tree filtering module that leverages the structural property of minimal spanning tree to model long-range dependencies while preserving the details. Furthermore, we propose a highly efficient linear-time algorithm to reduce resource consumption. Thus, the designed modules can be plugged into existing deep neural networks conveniently. To this end, tree filtering modules are embedded to formulate a unified framework for semantic segmentation. We conduct extensive ablation studies to elaborate on the effectiveness and efficiency of the proposed method. Specifically, it attains better performance with much less overhead compared with the classic PSP block and Non-local operation under the same backbone. Our approach is proved to achieve consistent improvements on several benchmarks without bells-and-whistles. Code and models are available at https://github.com/StevenGrove/TreeFilter-Torch.