Position: Evaluation of Visual Processing Should Be Human-Centered, Not Metric-Centered
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_ | 1866908871331676160 |
|---|---|
| author | Hu, Jinfan Yu, Fanghua You, Zhiyuan Yin, Xiang An, Hongyu Lin, Xinqi Dong, Chao Gu, Jinjin |
| author_facet | Hu, Jinfan Yu, Fanghua You, Zhiyuan Yin, Xiang An, Hongyu Lin, Xinqi Dong, Chao Gu, Jinjin |
| contents | This position paper argues that the evaluation of modern visual processing systems should no longer be driven primarily by single-metric image quality assessment benchmarks, particularly in the era of generative and perception-oriented methods. Image restoration exemplifies this divergence: while objective IQA metrics enable reproducible, scalable evaluation, they have increasingly drifted apart from human perception and user preferences. We contend that this mismatch risks constraining innovation and misguiding research progress across visual processing tasks. Rather than rejecting metrics altogether, this paper calls for a rebalancing of evaluation paradigms, advocating a more human-centered, context-aware, and fine-grained approach to assessing the visual models' outcomes. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_00643 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Position: Evaluation of Visual Processing Should Be Human-Centered, Not Metric-Centered Hu, Jinfan Yu, Fanghua You, Zhiyuan Yin, Xiang An, Hongyu Lin, Xinqi Dong, Chao Gu, Jinjin Computer Vision and Pattern Recognition This position paper argues that the evaluation of modern visual processing systems should no longer be driven primarily by single-metric image quality assessment benchmarks, particularly in the era of generative and perception-oriented methods. Image restoration exemplifies this divergence: while objective IQA metrics enable reproducible, scalable evaluation, they have increasingly drifted apart from human perception and user preferences. We contend that this mismatch risks constraining innovation and misguiding research progress across visual processing tasks. Rather than rejecting metrics altogether, this paper calls for a rebalancing of evaluation paradigms, advocating a more human-centered, context-aware, and fine-grained approach to assessing the visual models' outcomes. |
| title | Position: Evaluation of Visual Processing Should Be Human-Centered, Not Metric-Centered |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2603.00643 |