ATTA: Anomaly-aware Test-Time Adaptation for Out-of-Distribution Detection in Segmentation

TL;DR

This paper introduces ATTA, a dual-level test-time adaptation framework that improves out-of-distribution detection in segmentation tasks by jointly addressing domain and semantic shifts, leading to more accurate detection in real-world scenarios.

Contribution

The paper proposes a novel dual-level OOD detection method that distinguishes domain and semantic shifts and adaptively improves segmentation models under domain shift conditions.

Findings

Consistent performance improvements on OOD segmentation benchmarks.

Effective detection of domain and semantic shifts in diverse scenarios.

Enhanced model capacity for identifying novel classes.

Abstract

Recent advancements in dense out-of-distribution (OOD) detection have primarily focused on scenarios where the training and testing datasets share a similar domain, with the assumption that no domain shift exists between them. However, in real-world situations, domain shift often exits and significantly affects the accuracy of existing out-of-distribution (OOD) detection models. In this work, we propose a dual-level OOD detection framework to handle domain shift and semantic shift jointly. The first level distinguishes whether domain shift exists in the image by leveraging global low-level features, while the second level identifies pixels with semantic shift by utilizing dense high-level feature maps. In this way, we can selectively adapt the model to unseen domains as well as enhance model's capacity in detecting novel classes. We validate the efficacy of our proposed method on several OOD segmentation benchmarks, including those with significant domain shifts and those without, observing consistent performance improvements across various baseline models. Code is available at ${\href{https://github.com/gaozhitong/ATTA}{https://github.com/gaozhitong/ATTA}}$.