Reproduce “Hallucinated-IQA: No-Reference Image Quality Assessment via Adversarial Learning” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Kwan-Yee Lin, Guanxiang Wang
Hallucinated-IQA estimates image quality without an original reference at assessment time by generating a stand-in from the distorted image, then giving both and their discrepancy to a quality regressor. Training uses clean references and quality scores. On TID2013 across all 24 distortion types, it reaches 0.879 SROCC versus RankIQA’s 0.780, but performs less well on a small number of types.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.