MetaRefine improves image quality scores using original geometry and features
MetaRefine: Score Refinement with Multi-Level Features for Image Assessment
Computer Vision and Pattern RecognitionArtificial Intelligence
Summary
Images have important details like their shape and layout that affect how good they look or how technically sound they are. Many existing methods change these details by resizing or cropping images before judging their quality, which can hide important clues. The authors created MetaRefine, a system that adjusts image quality scores by combining original image shape information with multiple levels of visual details, without changing existing scoring methods. Their tests on many datasets show MetaRefine makes quality scores more accurate while adding very little extra computation.
What this means in practice
- •For photo editing software teams: Improve automated quality scoring of images by incorporating original image dimensions and detailed visual features to refine aesthetic and technical assessments.
- •For digital asset management teams: Enhance image quality evaluation to better filter or prioritize images using refined scoring that accounts for original geometry and multi-level visual cues.
Authors
Chuanzhi Xu, Chengkun Yue, Haolan Guo, Xuanhua Yin, Boyu Wei, Zihan Deng, Qiang Qu, Weidong Cai
Abstract
An image's original dimensions, aspect ratio, and composition provide important cues that can affect its technical and aesthetic quality. Image assessment methods use resizing, cropping, padding, or patch sampling to accommodate different input sizes. However, these operations can alter or obscure original geometric and compositional cues. Existing assessment models also encode rich multi-level visual evidence that can be further extracted and exploited to improve existing models. This motivates us to combine visual evidence with original-image geometry for score refinement without changing the original scoring pipelines. In this paper, we introduce MetaRefine, a metadata-aware score refinement architecture whose lightweight residual heads predict score corrections from geometry, multi-level visual features, and the frozen model's original score. Across 100 configurations on 21 datasets, MetaRefine improves PLCC in 88 configurations and SRCC in 85, with average relative gains of 7.55% and 7.10%, respectively. It reduces MAE in 90 configurations and RMSE in 89, with average relative reductions of 23.27% and 21.93%, respectively. The median parameter overhead is only 1.11%. MetaRefine can further improve state-of-the-art (SOTA) assessment models on multiple benchmarks.