Q-Ponder: A Unified Training Pipeline for Reasoning-based Visual Quality Assessment
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866908405548974080 |
|---|---|
| author | Cai, Zhuoxuan Zhang, Jian Yuan, Xinbin Jiang, Peng-Tao Chen, Wenxiang Tang, Bowen Yao, Lujian Wang, Qiyuan Chen, Jinwen Li, Bo |
| author_facet | Cai, Zhuoxuan Zhang, Jian Yuan, Xinbin Jiang, Peng-Tao Chen, Wenxiang Tang, Bowen Yao, Lujian Wang, Qiyuan Chen, Jinwen Li, Bo |
| contents | Recent studies demonstrate that multimodal large language models (MLLMs) can proficiently evaluate visual quality through interpretable assessments. However, existing approaches typically treat quality scoring and reasoning descriptions as separate tasks with disjoint optimization objectives, leading to a trade-off: models adept at quality reasoning descriptions struggle with precise score regression, while score-focused models lack interpretability. This limitation hinders the full potential of MLLMs in visual quality assessment, where accuracy and interpretability should be mutually reinforcing. To address this, we propose a unified two-stage training framework comprising a cold-start stage and a reinforcement learning-based fine-tuning stage. Specifically, in the first stage, we distill high-quality data from a teacher model through expert-designed prompts, initializing reasoning capabilities via cross-entropy loss supervision. In the second stage, we introduce a novel reward with Group Relative Policy Optimization (GRPO) to jointly optimize scoring accuracy and reasoning consistency. We designate the models derived from these two stages as Q-Ponder-CI and Q-Ponder. Extensive experiments show that Q-Ponder achieves state-of-the-art (SOTA) performance on quality score regression benchmarks, delivering up to 6.5% higher SRCC on cross-domain datasets. Furthermore, Q-Ponder significantly outperforms description-based SOTA models, including its teacher model Qwen-2.5-VL-72B, particularly in description accuracy and reasonableness, demonstrating the generalization potential over diverse tasks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_05384 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Q-Ponder: A Unified Training Pipeline for Reasoning-based Visual Quality Assessment Cai, Zhuoxuan Zhang, Jian Yuan, Xinbin Jiang, Peng-Tao Chen, Wenxiang Tang, Bowen Yao, Lujian Wang, Qiyuan Chen, Jinwen Li, Bo Computer Vision and Pattern Recognition Artificial Intelligence Multimedia Recent studies demonstrate that multimodal large language models (MLLMs) can proficiently evaluate visual quality through interpretable assessments. However, existing approaches typically treat quality scoring and reasoning descriptions as separate tasks with disjoint optimization objectives, leading to a trade-off: models adept at quality reasoning descriptions struggle with precise score regression, while score-focused models lack interpretability. This limitation hinders the full potential of MLLMs in visual quality assessment, where accuracy and interpretability should be mutually reinforcing. To address this, we propose a unified two-stage training framework comprising a cold-start stage and a reinforcement learning-based fine-tuning stage. Specifically, in the first stage, we distill high-quality data from a teacher model through expert-designed prompts, initializing reasoning capabilities via cross-entropy loss supervision. In the second stage, we introduce a novel reward with Group Relative Policy Optimization (GRPO) to jointly optimize scoring accuracy and reasoning consistency. We designate the models derived from these two stages as Q-Ponder-CI and Q-Ponder. Extensive experiments show that Q-Ponder achieves state-of-the-art (SOTA) performance on quality score regression benchmarks, delivering up to 6.5% higher SRCC on cross-domain datasets. Furthermore, Q-Ponder significantly outperforms description-based SOTA models, including its teacher model Qwen-2.5-VL-72B, particularly in description accuracy and reasonableness, demonstrating the generalization potential over diverse tasks. |
| title | Q-Ponder: A Unified Training Pipeline for Reasoning-based Visual Quality Assessment |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence Multimedia |
| url | https://arxiv.org/abs/2506.05384 |