用多尺度压缩分析检测图像伪造，无需GPU也能快速准确识别。
Format-Controlled Multi-Scale JPEG Compression Response Analysis for Image-Level Forgery Screening
- 基于七种质量的JPEG压缩误差分析，结合跨质量比特征捕捉拼接痕迹。
- 在去格式混淆的数据集上达到0.990 AUC和0.905 F1，性能接近完美。
- 完全基于CPU运行，特征可解释，适合资源受限场景使用。
图像伪造检测是数字取证的关键任务，但许多深度学习定位方法计算开销大且依赖GPU。本文提出一种轻量级、可解释的特征工程流程，仅使用CPU计算和梯度提升树进行图像级伪造筛查。方法引入七种JPEG质量级别的多尺度误差等级分析（ELA），并设计新颖的跨质量ELA比值特征，捕捉拼接区域特有的二次压缩痕迹，同时融合空间熵、FFT能量带、边缘密度、SRM残差和DCT块效应，生成405维特征向量。原数据集CASIA v2.0存在格式混杂问题（60%伪造图是TIFF而真实图是JPEG/BMP），导致一个简单的is_tiff分类器即可达0.80 AUC。为此，我们采用严格的格式控制评估：在仅含JPEG的子集（9,501张图像）上，本方法在5折分层交叉验证下获得AUC=0.990（95%置信区间：0.988–0.991）和F1=0.905；在保守的源感知组划分下，AUC仍保持0.976。消融实验表明，多尺度ELA贡献最大（相比单质量提升0.180 AUC），跨质量比值则提供互补的二次压缩检测能力。结果表明该方法真正检测的是压缩历史不一致而非文件格式捷径——兼具特征可解释性、纯CPU部署与亚秒级推理速度。
原文摘要 · Abstract (English)
Image forgery detection is a critical task in digital forensics, yet many deep-learning localization approaches are typically GPU-accelerated and computationally heavier than handcrafted screening methods. We propose a lightweight, interpretable feature engineering pipeline for image-level forgery screening using only CPU computation and gradient boosted trees. Our method introduces \emph{multi-scale Error Level Analysis} (ELA) computed at seven JPEG quality levels, combined with novel \emph{cross-quality ELA ratio} features that capture double-compression artifacts characteristic of spliced regions, augmented by spatial entropy, FFT energy bands, edge density, SRM residuals, and DCT blockiness, yielding a 405-dimensional feature vector. CASIA v2.0 contains a format confound (60\% of tampered images are TIFF while authentic images are JPEG/BMP and contain no TIFF samples), enabling a trivial \texttt{is\_tiff} classifier to reach 0.80 AUC. We address this through rigorous format-controlled evaluation: on the JPEG-only subset (9,501 images, eliminating the TIFF/JPEG container confound), our method achieves AUC~=~0.990 [95\% CI: 0.988--0.991] and F1~=~0.905 using 5-fold stratified cross-validation. Under a conservative source-aware group split (preventing related images from appearing in both train and test), AUC remains 0.976. An ablation study reveals that multi-scale ELA provides the dominant gain (+0.180 AUC over single-quality on the format-controlled subset), while cross-quality ratios provide complementary double-compression detection. These results support that the method detects compression-history inconsistencies rather than file-format shortcuts -- while offering feature-level interpretability, CPU-only deployment, and sub-second inference.
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。