BPCLIP propagates low-level distortion features upward through cross-attention and scores images by similarity to fixed CLIP text adjectives, reporting small gains on FR and NR IQA benchmarks.
Learning generalizable perceptual representations for data-efficient no-reference image quality assessment,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
baseline 1
citation-polarity summary
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1roles
baseline 1polarities
baseline 1representative citing papers
citing papers explorer
-
BPCLIP: A Bottom-up Image Quality Assessment from Distortion to Semantics Based on CLIP
BPCLIP propagates low-level distortion features upward through cross-attention and scores images by similarity to fixed CLIP text adjectives, reporting small gains on FR and NR IQA benchmarks.