Enhancing Multi-view Open-set Learning via Ambiguity Uncertainty Calibration and View-wise Debiasing

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Fang, Zihan, Xu, Zhiyong, Du, Lan, Du, Shide, Cai, Zhiling, Wang, Shiping
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866908479643451392
author Fang, Zihan
Xu, Zhiyong
Du, Lan
Du, Shide
Cai, Zhiling
Wang, Shiping
author_facet Fang, Zihan
Xu, Zhiyong
Du, Lan
Du, Shide
Cai, Zhiling
Wang, Shiping
contents Existing multi-view learning models struggle in open-set scenarios due to their implicit assumption of class completeness. Moreover, static view-induced biases, which arise from spurious view-label associations formed during training, further degrade their ability to recognize unknown categories. In this paper, we propose a multi-view open-set learning framework via ambiguity uncertainty calibration and view-wise debiasing. To simulate ambiguous samples, we design O-Mix, a novel synthesis strategy to generate virtual samples with calibrated open-set ambiguity uncertainty. These samples are further processed by an auxiliary ambiguity perception network that captures atypical patterns for improved open-set adaptation. Furthermore, we incorporate an HSIC-based contrastive debiasing module that enforces independence between view-specific ambiguous and view-consistent representations, encouraging the model to learn generalizable features. Extensive experiments on diverse multi-view benchmarks demonstrate that the proposed framework consistently enhances unknown-class recognition while preserving strong closed-set performance.
format Preprint
id arxiv_https___arxiv_org_abs_2508_01227
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Enhancing Multi-view Open-set Learning via Ambiguity Uncertainty Calibration and View-wise Debiasing
Fang, Zihan
Xu, Zhiyong
Du, Lan
Du, Shide
Cai, Zhiling
Wang, Shiping
Computer Vision and Pattern Recognition
Machine Learning
Existing multi-view learning models struggle in open-set scenarios due to their implicit assumption of class completeness. Moreover, static view-induced biases, which arise from spurious view-label associations formed during training, further degrade their ability to recognize unknown categories. In this paper, we propose a multi-view open-set learning framework via ambiguity uncertainty calibration and view-wise debiasing. To simulate ambiguous samples, we design O-Mix, a novel synthesis strategy to generate virtual samples with calibrated open-set ambiguity uncertainty. These samples are further processed by an auxiliary ambiguity perception network that captures atypical patterns for improved open-set adaptation. Furthermore, we incorporate an HSIC-based contrastive debiasing module that enforces independence between view-specific ambiguous and view-consistent representations, encouraging the model to learn generalizable features. Extensive experiments on diverse multi-view benchmarks demonstrate that the proposed framework consistently enhances unknown-class recognition while preserving strong closed-set performance.
title Enhancing Multi-view Open-set Learning via Ambiguity Uncertainty Calibration and View-wise Debiasing
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2508.01227