Official Repo for Revisiting Vision–Language Foundations for No-Reference Image Quality Assessment
-
Updated
May 14, 2026 - Python
Official Repo for Revisiting Vision–Language Foundations for No-Reference Image Quality Assessment
A tagged catalog of perceptual quality assessment: image quality, video quality, UGC and HDR. Methods and subjective datasets in one place, generated from CSV.
Retrieval-augmented NR-IQA: 5-encoder fusion + closed-form regression, AutoML model selection across 6 IQA benchmarks
Content-variant reference IQA via knowledge distillation: a full-reference teacher transfers high-quality distribution priors to a student that needs only a non-pixel-aligned / content-variant reference, so quality scores no longer require a pixel-perfect reference image.
Official PyTorch implementation of CaliTrace, monotone spatial-penalty routing for AI-generated image quality assessment.
No-reference CT image quality assessment using a RadImageNet-pretrained ResNet50, adapted from the Ohashi et al. methodology to the LDCT-IQAC dataset with real human quality scores.
End-to-end image-quality regression & measurement platform: provenance catalog, objective + learned no-reference perceptual metrics calibrated against human MOS, build-over-build regression detection, FastAPI dashboard + natural-language query.
To associate your repository with the no-reference-iqa topic, visit your repo's landing page and select "manage topics."