Research map / Video / Video segmentation and tracking
Generalizable Video Quality Assessment via Weak-to-Strong Learning
Linhan Cao, Wei Sun, Xiangyang Zhu, Kaiwei Zhang, Jun Jia, Yicong Peng, Dandan Zhu, Guangtao Zhai, Xiongkuo Min
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add homogeneous teacher ensembling (ablation) | W2S student · OOD overall | SRCC: 0.591 -> 0.602 | Improved |
| add synthetic distortion heterogeneous teachers (ablation) | W2S student · OOD overall | SRCC: 0.602 -> 0.650 | Improved |
| add confidence loss (ablation) | W2S student · OOD overall | SRCC: 0.650 -> 0.672 | Improved |
| add iterative stage 2 W2S training (ablation) | W2S student · OOD overall | SRCC: 0.672 -> 0.722 | Improved |
| add iterative stage 3 W2S training (ablation) | W2S student · OOD overall | SRCC: 0.722 -> 0.745 | Improved |
| add difficult-sample selection in stage 2 (ablation) | W2S model (V) · OOD overall | SRCC: 0.669 -> 0.722 | Improved |
| add pseudo-label refinement in stage 2 (ablation) | W2S model (V) · OOD overall | SRCC: 0.657 -> 0.722 | Improved |
| add heterogeneous teachers (ablation) | W2S student · In-domain overall | SRCC: 0.856 -> 0.858 | No clear effect |
| train student on weak teacher labels instead of teacher itself | MinimalisticVQA(VII) to student (27k) · OOD overall | SRCC: 0.490 -> 0.515 | Improved |
| train student on weak teacher labels | student (27k) vs teacher · In-domain overall | SRCC | No clear effect |
| scale student training videos from 27k to 200k | MinimalisticVQA(IX)-labeled student · OOD overall | SRCC: 0.582 -> 0.591 | Improved |
| replace LSVQ ground-truth labels with weak labels | MinimalisticVQA(IX) student (27k) · OOD overall | SRCC: 0.577 -> 0.582 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- VQAThinker — VQAThinker: Exploring Generalizable and Explainable Video Quality Assessment via Reinforcement Learning (AAAI 2026) · better on 4 of 4 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- VQAThinker: Exploring Generalizable and Explainable Video Quality Assessment via Reinforcement Learning (AAAI 2026 · similarity 0.91)
- MDS-VQA: Model-Informed Data Selection for Video Quality Assessment (CVPR 2026 · similarity 0.91)
- Knowledge Guided Semi-supervised Learning for Quality Assessment of User Generated Videos (AAAI 2024 · similarity 0.88)
- PTM-VQA: Efficient Video Quality Assessment Leveraging Diverse PreTrained Models from the Wild (CVPR 2024 · similarity 0.88)
- Q-Bench-Video: Benchmark the Video Quality Understanding of LMMs (CVPR 2025 · similarity 0.87)
- KVQ: Boosting Video Quality Assessment via Saliency-guided Local Perception (CVPR 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
