AI Paper Atlas

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

CVPR 2026 · Video › Video segmentation and tracking · approach: Video restoration with temporal priors · 7 citations (OpenAlex)

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 changedWhereMeasuredEffect
add homogeneous teacher ensembling (ablation)W2S student · OOD overallSRCC: 0.591 -> 0.602Improved
add synthetic distortion heterogeneous teachers (ablation)W2S student · OOD overallSRCC: 0.602 -> 0.650Improved
add confidence loss (ablation)W2S student · OOD overallSRCC: 0.650 -> 0.672Improved
add iterative stage 2 W2S training (ablation)W2S student · OOD overallSRCC: 0.672 -> 0.722Improved
add iterative stage 3 W2S training (ablation)W2S student · OOD overallSRCC: 0.722 -> 0.745Improved
add difficult-sample selection in stage 2 (ablation)W2S model (V) · OOD overallSRCC: 0.669 -> 0.722Improved
add pseudo-label refinement in stage 2 (ablation)W2S model (V) · OOD overallSRCC: 0.657 -> 0.722Improved
add heterogeneous teachers (ablation)W2S student · In-domain overallSRCC: 0.856 -> 0.858No clear effect
train student on weak teacher labels instead of teacher itselfMinimalisticVQA(VII) to student (27k) · OOD overallSRCC: 0.490 -> 0.515Improved
train student on weak teacher labelsstudent (27k) vs teacher · In-domain overallSRCCNo clear effect
scale student training videos from 27k to 200kMinimalisticVQA(IX)-labeled student · OOD overallSRCC: 0.582 -> 0.591Improved
replace LSVQ ground-truth labels with weak labelsMinimalisticVQA(IX) student (27k) · OOD overallSRCC: 0.577 -> 0.582Improved

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.

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

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy