Anomaly Segmentation Using Class-aware Erosion and Smoothing

Kang, Beoungwoo; Kwak, Jaeho; Kang, Suk‐Ju

doi:10.1109/icce-asia57006.2022.9954841

Cited by 1 publication

(1 citation statement)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The existing methods for such tasks can be distinguished by whether they use OOD data for training. Some methods expand the training set to include OOD images from other datasets (Hendrycks, Mazeika, and Dietterich 2018a;Chan, Rottmann, and Gottschalk 2021a;Kang, Kwak, and Kang 2022;Tian et al 2022), or utilize large-scale models, e.g., SAM (Kirillov et al 2023) to generate region proposals. Such expansion-based approaches are not of great interest in this paper since we aim to solve the general problem of OOD segmentation without having access to any OOD images for training.…”

Section: Introductionmentioning

confidence: 99%

CSL: Class-Agnostic Structure-Constrained Learning for Segmentation Including the Unseen

Zhang,

Li,

et al. 2024

AAAI

View full text Add to dashboard Cite

Addressing Out-Of-Distribution (OOD) Segmentation and Zero-Shot Semantic Segmentation (ZS3) is challenging, necessitating segmenting unseen classes. Existing strategies adapt the class-agnostic Mask2Former (CA-M2F) tailored to specific tasks. However, these methods cater to singular tasks, demand training from scratch, and we demonstrate certain deficiencies in CA-M2F, which affect performance. We propose the Class-Agnostic Structure-Constrained Learning (CSL), a plug-in framework that can integrate with existing methods, thereby embedding structural constraints and achieving performance gain, including the unseen, specifically OOD, ZS3, and domain adaptation (DA) tasks. There are two schemes for CSL to integrate with existing methods (1) by distilling knowledge from a base teacher network, enforcing constraints across training and inference phrases, or (2) by leveraging established models to obtain per-pixel distributions without retraining, appending constraints during the inference phase. Our soft assignment and mask split methodologies enhance OOD object segmentation. Empirical evaluations demonstrate CSL's prowess in boosting the performance of existing algorithms spanning OOD segmentation, ZS3, and DA segmentation, consistently transcending the state-of-art across all three tasks.

show abstract

Section: Introductionmentioning

confidence: 99%