让图像质量评估模型更敏感地捕捉局部退化，提升真实场景适应性。
Spatially Localized Image Degradation Embeddings for Image Quality Assessment
- 设计双分支ViT架构，引入局部退化增强对比学习。
- 提出阈值限定排除机制，解决局部退化带来的结构冲突。
- 仅用合成数据预训练，仍达先进水平且对局部退化更敏感。
自监督学习（SSL）目前主导无参考图像质量评估（NR-IQA）的前沿性能。然而，现有标准SSL流程对整幅图像统一施加合成退化，难以捕捉真实内容中常见的局部且共现的退化特征。本文实证揭示了现有顶尖编码器在空间局部退化上的表征盲区，其敏感性显著下降。为此，我们提出空间局部化图像退化嵌入（SLIDE-IQA），采用双分支视觉变换器框架，将空间受限的退化注入对比学习目标。为应对此类退化的空间复杂性，提出阈值限定排除机制（Threshold-Bounded Exclusion Mechanism），通过表征设计化解局部退化引发的结构冲突，确保潜在空间同时保留退化类型与空间尺度信息。最终，仅使用合成数据预训练的SLIDE-IQA，在保持对局部退化高敏感性的同时，在多个主流NR-IQA基准上达到与现有先进模型相当的性能。
原文摘要 · Abstract (English)
Self-supervised learning (SSL) currently drives state-of-the-art performance in no-reference image quality assessment (NR-IQA). However, standard SSL pipelines uniformly apply synthetic distortions across the entire image field, which can limit their sensitivity to spatially localized and co-occurring degradations encountered in real-world content. In this work, we empirically expose this representational blind spot across existing state-of-the-art encoders, demonstrating their reduced sensitivity to spatially bounded image degradations. To bridge this gap, we introduce Spatial Localized Image Degradation Embeddings for Image Quality Assessment (SLIDE-IQA). SLIDE-IQA employs a dual-branch Vision Transformer framework that injects spatially bounded degradations into a contrastive pretraining objective. To handle the spatial complexity of these degradations, we introduce a Threshold-Bounded Exclusion Mechanism, a representational design choice that resolves structural conflicts arising from spatially localized distortions to ensure the latent space respects both degradation type and spatial scale. Finally, we show that SLIDE-IQA's synthetic-only pretraining significantly improves sensitivity to localized distortions, while achieving competitive performance on NR-IQA benchmarks against existing SSL NR-IQA models.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。