Lightweight Frequency Masker for Cross-Domain Few-Shot Semantic Segmentation

Jintao Tong, Yixiong Zou, Yuhua Li, Ruixuan Li

Keywords: cross-domain few-shot segmentation, frequency component, feature disentanglement，channel attention

TL;DR: We find simply filtering frequency components significantly improves CD-FSS performance. We delve into this for an interpretation, and further propose a lightweight frequency masker, importing only 2.5% parameters but averagely improving over 11%.

Abstract: Cross-domain few-shot segmentation (CD-FSS) is proposed to first pre-train the model on a large-scale source-domain dataset, and then transfer the model to data-scarce target-domain datasets for pixel-level segmentation. The significant domain gap between the source and target datasets leads to a sharp decline in the performance of existing few-shot segmentation (FSS) methods in cross-domain scenarios. In this work, we discover an intriguing phenomenon: simply filtering different frequency components for target domains can lead to a significant performance improvement, sometimes even as high as 14% mIoU. Then, we delve into this phenomenon for an interpretation, and find such improvements stem from the reduced inter-channel correlation in feature maps, which benefits CD-FSS with enhanced robustness against domain gaps and larger activated regions for segmentation. Based on this, we propose a lightweight frequency masker, which further reduces channel correlations by an amplitude-phase-masker (APM) module and an Adaptive Channel Phase Attention (ACPA) module. Notably, APM introduces only 0.01% additional parameters but improves the average performance by over 10%, and ACPA imports only 2.5% parameters but further improves the performance by over 1.5%, which significantly surpasses the state-of-the-art CD-FSS methods.

Primary Area: Machine vision

Submission Number: 6768