Single Image Test-Time Adaptation for Segmentation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Janouskova, Klara, Shor, Tamir, Baskin, Chaim, Matas, Jiri
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911941392334848
author Janouskova, Klara
Shor, Tamir
Baskin, Chaim
Matas, Jiri
author_facet Janouskova, Klara
Shor, Tamir
Baskin, Chaim
Matas, Jiri
contents Test-Time Adaptation (TTA) methods improve the robustness of deep neural networks to domain shift on a variety of tasks such as image classification or segmentation. This work explores adapting segmentation models to a single unlabelled image with no other data available at test-time. In particular, this work focuses on adaptation by optimizing self-supervised losses at test-time. Multiple baselines based on different principles are evaluated under diverse conditions and a novel adversarial training is introduced for adaptation with mask refinement. Our additions to the baselines result in a 3.51 and 3.28 % increase over non-adapted baselines, without these improvements, the increase would be 1.7 and 2.16 % only.
format Preprint
id arxiv_https___arxiv_org_abs_2309_14052
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Single Image Test-Time Adaptation for Segmentation
Janouskova, Klara
Shor, Tamir
Baskin, Chaim
Matas, Jiri
Computer Vision and Pattern Recognition
Test-Time Adaptation (TTA) methods improve the robustness of deep neural networks to domain shift on a variety of tasks such as image classification or segmentation. This work explores adapting segmentation models to a single unlabelled image with no other data available at test-time. In particular, this work focuses on adaptation by optimizing self-supervised losses at test-time. Multiple baselines based on different principles are evaluated under diverse conditions and a novel adversarial training is introduced for adaptation with mask refinement. Our additions to the baselines result in a 3.51 and 3.28 % increase over non-adapted baselines, without these improvements, the increase would be 1.7 and 2.16 % only.
title Single Image Test-Time Adaptation for Segmentation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2309.14052