A Causally Grounded Taxonomy for Image Degradation Robustness Evaluation

2026-05-15

AI Summary

Image degradation is pervasive throughout the imaging pipeline, yet existing research lacks a unified taxonomy and evaluation protocol, hindering cross-dataset and cross-task comparisons. This work introduces a causal perspective to address this gap, proposing a dual-axis classification framework: one axis categorizes degradations by their dominant causal source in the imaging pipeline—encompassing environment, sensor/optics, ISP/codec, and transmission systems—while the other characterizes their perceptual effects, augmented with a lightweight severity quantification layer. Built upon this framework, the COCO Degradation benchmark leverages PSNR, SSIM, and LPIPS to uniformly measure degradation intensity across physical artifacts, algorithmic perturbations, and perceptual distortions, substantially enhancing the evaluation of object detection model robustness under diverse imaging conditions.

Abstract

Image degradations can occur during acquisition, processing, and transmission, altering visual appearance and affecting downstream vision tasks. They are studied in several communities, including synthetic corruption benchmarks for robustness evaluation, perceptual image quality assessment, and physically grounded analyses of imaging systems or real camera failures. Although these areas address closely related phenomena, they often use incompatible grouping schemes and backend specific severity definitions, making results difficult to compare across datasets, degradation sources, and tasks. We propose a causally grounded framework for organizing and interpreting image degradations across these settings. Instead of introducing new degradations or redefining existing benchmarks, we provide an interpretive representation and measurement layer that makes implicit assumptions explicit. Each degradation is described along two orthogonal axes: its dominant causal source in the imaging pipeline (environment, sensor/optics, ISP/renderer/codec, or transfer/system), and its resulting perceptual effect. This dual axis abstraction yields a compact taxonomy spanning algorithmic corruptions, perceptual distortions, and physically motivated imaging artifacts. To address inconsistent severity semantics without changing existing implementations, we introduce a lightweight severity measurement layer. For every degradation and each native severity level of a given backend, we quantify degradation strength using full reference image quality metrics: PSNR, SSIM, and LPIPS. This makes severity observable and comparable across sources while preserving native parameterizations. We demonstrate the framework through COCO Degradation, a taxonomy aligned benchmark for evaluating object detector robustness under diverse imaging conditions.

Problem

• image degradation
• robustness evaluation
• taxonomy
• severity semantics
• cross-dataset comparison

Innovation

• causal taxonomy
• image degradation
• robustness evaluation
• severity normalization
• perceptual effects