VersusQ: Pairwise Margin Reasoning for Generalizable Video Quality Assessment
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_ | 1866918514419302400 |
|---|---|
| author | Meng, Shibei Yang, Binxin Liu, Yuan Zhang, Jiexuan Lv, Zhengyao Yin, Hubery Xu, Qiang |
| author_facet | Meng, Shibei Yang, Binxin Liu, Yuan Zhang, Jiexuan Lv, Zhengyao Yin, Hubery Xu, Qiang |
| contents | Large Multimodal Models (LMMs) have shown promise for video quality assessment, but most methods still predict an absolute score for each video. Such pointwise supervision often mixes perceptual quality with dataset-specific calibration, including annotation protocols, rating habits, and score distributions. As a result, the learned scoring rule may work well within a benchmark but transfer poorly across unseen domains. We argue that relative comparisons alleviate the absolute-scale calibration bias by focusing purely on perceptual differences rather than dataset-specific rating habits. Consequently, we propose \textbf{VersusQ}, a pairwise margin reasoning framework driven entirely by direct comparisons. Specifically, VersusQ performs LMM-based comparison between two videos, reasons about their visual and temporal quality differences, and predicts a signed continuous margin that captures both the preferred choice and the degree of difference. Furthermore, to align interpretable comparison rationales with fine-grained numerical differences, we introduce Margin-Coupled GRPO, which jointly optimizes rollout-based relational reasoning and continuous margin regression. Extensive experiments on multiple public VQA benchmarks demonstrate that VersusQ achieves state-of-the-art performance, strong cross-domain generalization, and reliable fine-grained ranking under heterogeneous evaluation scenarios. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_21130 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | VersusQ: Pairwise Margin Reasoning for Generalizable Video Quality Assessment Meng, Shibei Yang, Binxin Liu, Yuan Zhang, Jiexuan Lv, Zhengyao Yin, Hubery Xu, Qiang Computer Vision and Pattern Recognition Large Multimodal Models (LMMs) have shown promise for video quality assessment, but most methods still predict an absolute score for each video. Such pointwise supervision often mixes perceptual quality with dataset-specific calibration, including annotation protocols, rating habits, and score distributions. As a result, the learned scoring rule may work well within a benchmark but transfer poorly across unseen domains. We argue that relative comparisons alleviate the absolute-scale calibration bias by focusing purely on perceptual differences rather than dataset-specific rating habits. Consequently, we propose \textbf{VersusQ}, a pairwise margin reasoning framework driven entirely by direct comparisons. Specifically, VersusQ performs LMM-based comparison between two videos, reasons about their visual and temporal quality differences, and predicts a signed continuous margin that captures both the preferred choice and the degree of difference. Furthermore, to align interpretable comparison rationales with fine-grained numerical differences, we introduce Margin-Coupled GRPO, which jointly optimizes rollout-based relational reasoning and continuous margin regression. Extensive experiments on multiple public VQA benchmarks demonstrate that VersusQ achieves state-of-the-art performance, strong cross-domain generalization, and reliable fine-grained ranking under heterogeneous evaluation scenarios. |
| title | VersusQ: Pairwise Margin Reasoning for Generalizable Video Quality Assessment |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2605.21130 |