When Relations Break: Analyzing Relation Hallucination in Vision-Language Model Under Rotation and Noise
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866915997251796992 |
|---|---|
| author | Shin, Philip Wootaek Sridhar, Ajay Narayanan Devarapalli, Sivani Zhang, Rui Sampson, Jack Narayanan, Vijaykrishnan |
| author_facet | Shin, Philip Wootaek Sridhar, Ajay Narayanan Devarapalli, Sivani Zhang, Rui Sampson, Jack Narayanan, Vijaykrishnan |
| contents | Vision-language models (VLMs) achieve strong multimodal performance but remain prone to relation hallucination, which requires accurate reasoning over inter-object interactions. We study the impact of visual perturbations, specifically rotation and noise, and show that even mild distortions significantly degrade relational reasoning across models and datasets. We further evaluate prompt-based augmentation and preprocessing strategies (orientation correction and denoising), finding that while they offer partial improvements, they do not fully resolve hallucinations. Our results reveal a gap between perceptual robustness and relational understanding, highlighting the need for more robust, geometry-aware VLMs. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_05045 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | When Relations Break: Analyzing Relation Hallucination in Vision-Language Model Under Rotation and Noise Shin, Philip Wootaek Sridhar, Ajay Narayanan Devarapalli, Sivani Zhang, Rui Sampson, Jack Narayanan, Vijaykrishnan Computer Vision and Pattern Recognition Computation and Language Vision-language models (VLMs) achieve strong multimodal performance but remain prone to relation hallucination, which requires accurate reasoning over inter-object interactions. We study the impact of visual perturbations, specifically rotation and noise, and show that even mild distortions significantly degrade relational reasoning across models and datasets. We further evaluate prompt-based augmentation and preprocessing strategies (orientation correction and denoising), finding that while they offer partial improvements, they do not fully resolve hallucinations. Our results reveal a gap between perceptual robustness and relational understanding, highlighting the need for more robust, geometry-aware VLMs. |
| title | When Relations Break: Analyzing Relation Hallucination in Vision-Language Model Under Rotation and Noise |
| topic | Computer Vision and Pattern Recognition Computation and Language |
| url | https://arxiv.org/abs/2605.05045 |