Proposes GLIA framework to adapt Vision Transformers for blind image quality assessment via dual-stream global-local interaction, claiming higher accuracy and robustness with reduced parameters.
Massive online crowdsourced study of subjective and objective picture quality
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Unleashing Vision Transformer Potential In Image Quality Assessment via Global-Local Adaptive Interaction
Proposes GLIA framework to adapt Vision Transformers for blind image quality assessment via dual-stream global-local interaction, claiming higher accuracy and robustness with reduced parameters.