DCAC: Dynamic Class-Aware Cache Creates Stronger Out-of-Distribution Detectors
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866915737743917056 |
|---|---|
| author | Wu, Yanqi Chen, Qichao Lai, Runhe Lu, Xinhua Zhuang, Jia-Xin Zhao, Zhilin Zheng, Wei-Shi Wang, Ruixuan |
| author_facet | Wu, Yanqi Chen, Qichao Lai, Runhe Lu, Xinhua Zhuang, Jia-Xin Zhao, Zhilin Zheng, Wei-Shi Wang, Ruixuan |
| contents | Out-of-distribution (OOD) detection remains a fundamental challenge for deep neural networks, particularly due to overconfident predictions on unseen OOD samples during testing. We reveal a key insight: OOD samples predicted as the same class, or given high probabilities for it, are visually more similar to each other than to the true in-distribution (ID) samples. Motivated by this class-specific observation, we propose DCAC (Dynamic Class-Aware Cache), a training-free, test-time calibration module that maintains separate caches for each ID class to collect high-entropy samples and calibrate the raw predictions of input samples. DCAC leverages cached visual features and predicted probabilities through a lightweight two-layer module to mitigate overconfident predictions on OOD samples. This module can be seamlessly integrated with various existing OOD detection methods across both unimodal and vision-language models while introducing minimal computational overhead. Extensive experiments on multiple OOD benchmarks demonstrate that DCAC significantly enhances existing methods, achieving substantial improvements, i.e., reducing FPR95 by 6.55% when integrated with ASH-S on ImageNet OOD benchmark. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_12468 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | DCAC: Dynamic Class-Aware Cache Creates Stronger Out-of-Distribution Detectors Wu, Yanqi Chen, Qichao Lai, Runhe Lu, Xinhua Zhuang, Jia-Xin Zhao, Zhilin Zheng, Wei-Shi Wang, Ruixuan Computer Vision and Pattern Recognition Out-of-distribution (OOD) detection remains a fundamental challenge for deep neural networks, particularly due to overconfident predictions on unseen OOD samples during testing. We reveal a key insight: OOD samples predicted as the same class, or given high probabilities for it, are visually more similar to each other than to the true in-distribution (ID) samples. Motivated by this class-specific observation, we propose DCAC (Dynamic Class-Aware Cache), a training-free, test-time calibration module that maintains separate caches for each ID class to collect high-entropy samples and calibrate the raw predictions of input samples. DCAC leverages cached visual features and predicted probabilities through a lightweight two-layer module to mitigate overconfident predictions on OOD samples. This module can be seamlessly integrated with various existing OOD detection methods across both unimodal and vision-language models while introducing minimal computational overhead. Extensive experiments on multiple OOD benchmarks demonstrate that DCAC significantly enhances existing methods, achieving substantial improvements, i.e., reducing FPR95 by 6.55% when integrated with ASH-S on ImageNet OOD benchmark. |
| title | DCAC: Dynamic Class-Aware Cache Creates Stronger Out-of-Distribution Detectors |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2601.12468 |