A depth-guided cross-attention network with a transformer-CNN bridge improves no-reference image quality prediction and cross-dataset generalization over several published baselines.
Deep Neural Net- works for No-Reference and Full-Reference Image Qual- ity Assessment
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
DGIQA: Depth-guided Feature Attention and Refinement for Generalizable Image Quality Assessment
A depth-guided cross-attention network with a transformer-CNN bridge improves no-reference image quality prediction and cross-dataset generalization over several published baselines.