Few-shot Semantic Segmentation with Support-induced Graph Convolutional Network
Jie Liu (University of Amsterdam),* Yanqi Bao (Northeastern University), Wenzhe Yin (University of Amsterdam), haochen wang (UvA), Yang Gao (Nanjing University), Jan-Jakob Sonke (The Netherlands Cancer Institute), Efstratios Gavves (University of Amsterdam )
The 33rd British Machine Vision Conference
Abstract
Few-shot semantic segmentation (FSS) aims to achieve novel objects segmentation with only a few annotated samples and has made great progress recently. Most of the existing FSS models focus on the feature matching between support and query to tackle FSS. However, the appearance variations between objects from the same category could be extremely large, leading to unreliable feature matching and query mask prediction. To this end, we propose a Support-induced Graph Convolutional Network (SiGCN) to explicitly excavate latent context structure in query images. Specifically, we propose a Support-induced Graph Reasoning (SiGR) module to capture salient query object parts at different semantic levels with a Support-induced GCN. Furthermore, an instance association (IA) module is designed to capture high-order instance context from both support and query instances. By integrating the proposed two modules, SiGCN is able to learn rich query context representation, thus being more robust to appearance variations. Extensive experiments on PASCAL-5i and COCO-20i demonstrate that our SiGCN achieves state-of-the-art performance.Video
Citation
@inproceedings{Liu_2022_BMVC,
author = {Jie Liu and Yanqi Bao and Wenzhe Yin and haochen wang and Yang Gao and Jan-Jakob Sonke and Efstratios Gavves},
title = {Few-shot Semantic Segmentation with Support-induced Graph Convolutional Network},
booktitle = {33rd British Machine Vision Conference 2022, {BMVC} 2022, London, UK, November 21-24, 2022},
publisher = {{BMVA} Press},
year = {2022},
url = {https://bmvc2022.mpi-inf.mpg.de/0126.pdf}
}
Copyright © 2022
The British Machine Vision Conference is organised by