Few-Shot Image Segmentation Based on Dual Comparison Module and Sequential k-Shot Integration
Few-shot image segmentation intends to segment query images (test images) given only a few support samples with annotations. However, previous works ignore the impact of the object scales, especially in the support images. Meanwhile, current models only work on images with the similar size of the ob...
Main Authors: | , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Springer
2021-03-01
|
Series: | International Journal of Computational Intelligence Systems |
Subjects: | |
Online Access: | https://www.atlantis-press.com/article/125953300/view |
_version_ | 1811301229275381760 |
---|---|
author | Chencong Xing Shujing Lyu Yue Lu |
author_facet | Chencong Xing Shujing Lyu Yue Lu |
author_sort | Chencong Xing |
collection | DOAJ |
description | Few-shot image segmentation intends to segment query images (test images) given only a few support samples with annotations. However, previous works ignore the impact of the object scales, especially in the support images. Meanwhile, current models only work on images with the similar size of the object and rarely test on other domains. This paper proposes a new few-shot segmentation model named DCNet, which fully exploits the support set images and their annotations and is able to generalize to the test images with unseen objects of various scales. The idea is to gradually compare the features from the query and the support image, and refine the features for the query. Furthermore, a sequential k-shot comparison method is proposed based on the ConvGRU to integrate features from multiple annotated support images. Experiments on Pascal VOC dataset and X-ray Security Images demonstrate the excellent generalization performance of our model. |
first_indexed | 2024-04-13T07:05:08Z |
format | Article |
id | doaj.art-17c577dd85ee4db39d586285e48ff14a |
institution | Directory Open Access Journal |
issn | 1875-6883 |
language | English |
last_indexed | 2024-04-13T07:05:08Z |
publishDate | 2021-03-01 |
publisher | Springer |
record_format | Article |
series | International Journal of Computational Intelligence Systems |
spelling | doaj.art-17c577dd85ee4db39d586285e48ff14a2022-12-22T02:57:00ZengSpringerInternational Journal of Computational Intelligence Systems1875-68832021-03-0114110.2991/ijcis.d.210212.003Few-Shot Image Segmentation Based on Dual Comparison Module and Sequential k-Shot IntegrationChencong XingShujing LyuYue LuFew-shot image segmentation intends to segment query images (test images) given only a few support samples with annotations. However, previous works ignore the impact of the object scales, especially in the support images. Meanwhile, current models only work on images with the similar size of the object and rarely test on other domains. This paper proposes a new few-shot segmentation model named DCNet, which fully exploits the support set images and their annotations and is able to generalize to the test images with unseen objects of various scales. The idea is to gradually compare the features from the query and the support image, and refine the features for the query. Furthermore, a sequential k-shot comparison method is proposed based on the ConvGRU to integrate features from multiple annotated support images. Experiments on Pascal VOC dataset and X-ray Security Images demonstrate the excellent generalization performance of our model.https://www.atlantis-press.com/article/125953300/viewFew-shot learningImage segmentationDual comparison moduleConvolutional-gated recurrent unit |
spellingShingle | Chencong Xing Shujing Lyu Yue Lu Few-Shot Image Segmentation Based on Dual Comparison Module and Sequential k-Shot Integration International Journal of Computational Intelligence Systems Few-shot learning Image segmentation Dual comparison module Convolutional-gated recurrent unit |
title | Few-Shot Image Segmentation Based on Dual Comparison Module and Sequential k-Shot Integration |
title_full | Few-Shot Image Segmentation Based on Dual Comparison Module and Sequential k-Shot Integration |
title_fullStr | Few-Shot Image Segmentation Based on Dual Comparison Module and Sequential k-Shot Integration |
title_full_unstemmed | Few-Shot Image Segmentation Based on Dual Comparison Module and Sequential k-Shot Integration |
title_short | Few-Shot Image Segmentation Based on Dual Comparison Module and Sequential k-Shot Integration |
title_sort | few shot image segmentation based on dual comparison module and sequential k shot integration |
topic | Few-shot learning Image segmentation Dual comparison module Convolutional-gated recurrent unit |
url | https://www.atlantis-press.com/article/125953300/view |
work_keys_str_mv | AT chencongxing fewshotimagesegmentationbasedondualcomparisonmoduleandsequentialkshotintegration AT shujinglyu fewshotimagesegmentationbasedondualcomparisonmoduleandsequentialkshotintegration AT yuelu fewshotimagesegmentationbasedondualcomparisonmoduleandsequentialkshotintegration |