Do Underwater Image Quality Metrics Agree with Human Perception? A Benchmark with Human Choices, Simple Edits and Generative Drift
November 12, 2026
Title: Do Underwater Image Quality Metrics Agree with Human Perception? A Benchmark with Human Choices, Simple Edits and Generative Drift
Source: huggingface
Evaluations of 73 no-reference image quality assessment metrics reveal significant misalignment between automated scores and human preference in underwater enhancement. The benchmark tests handcrafted, learned, and vision-language models against paired-comparison studies to identify metric sensitivity to adversarial perturbations and generative drift.