分离语义与失真，多尺度双流模型提升AI生成图像质量评估精度。
Decoupling Semantics from Distortions: Multi-Scale Two-Stream Vision-Language Alignment for AI-Generated Image Quality Assessment
- 双流结构分层处理语义与纹理，粗粒度与细粒度特征分别捕捉整体与细节
- 在五个基准上实现1.11%和2.35%的SRCC提升，参数仅0.8M
- 适合关注生成图像真实感、细节退化的研究者与工程师
现有基于视觉语言模型（VLM）的AI生成图像质量评估方法存在语义-失真维度冲突：单一表征优化于语义区分，导致构图理解与低层感知纠缠，难以察觉细微质量退化。本文提出MST-CLIPIQA，一种多尺度双流框架，通过显式表征解耦实现分层视觉-语言对齐。架构采用双CLIP编码器，分别以粗粒度与细粒度补丁粒度运行：粗流捕捉全局语义连贯性，细流保留纹理特征与伪影模式。基于信息瓶颈的门控融合机制实现自适应跨尺度提炼，支持在有生成提示时进行提示锚定的对应评估。在五个基准上的实验验证了新最佳性能，质量评估平均提升1.11% SRCC，文本-图像对应预测提升2.35% SRCC，且仅需0.8M可训练参数。项目代码已开源。
原文摘要 · Abstract (English)
Existing vision-language model (VLM)-based AI-generated image quality assessment (AIGIQA) methods suffer from a fundamental semantic-distortion dimensional conflict: monolithic representations optimized for semantic discrimination inherently entangle compositional understanding with low-level perceptual sensitivity, rendering them blind to fine-grained quality degradations. We introduce MST-CLIPIQA, a multi-scale two-stream framework that achieves hierarchical vision-language alignment through explicit representational decoupling. Our architecture leverages dual CLIP encoders with complementary patch granularities: coarse-grained streams capture global semantic coherence while fine-grained streams preserve textural signatures and artifact patterns. An information bottleneck-inspired gated fusion mechanism performs adaptive cross-scale distillation, with optional cross-attention enabling prompt-anchored correspondence evaluation when generation prompts are available. Extensive experiments across five benchmarks establish new state-of-the-art results, achieving average improvements of 1.11 percent SRCC on quality and 2.35 percent SRCC on text-image correspondence prediction, while maintaining efficiency with only 0.8M trainable parameters. Our project is available at https://github.com/YMlinfeng/MST-CLIPIQA.
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。