Efficient Document Tampering Localization with Multi-Level Discrepancy Features and Unified DCT-Quantization Embedding

2026-06-20

AI Summary

This work addresses the limited generalization of existing methods in cross-domain and human-crafted forgery scenarios, as well as their overreliance on synthetic data, by proposing DiffNet—a novel architecture that fuses RGB and DCT modalities at an early stage. DiffNet introduces a lightweight, multi-level difference transformation to emphasize manipulation-induced inconsistencies and innovatively incorporates a frequency-index-aware joint embedding mechanism that integrates DCT coefficients with quantization tables, thereby redefining the DCT backbone paradigm. Coupled with multi-scale zero-sum filters and an efficient decoder, DiffNet substantially enhances sensitivity to subtle tampering traces, achieving state-of-the-art performance in cross-domain and human-made document tampering localization—surpassing prior methods by approximately 30% in accuracy while accelerating inference throughput by up to 7×.

Abstract

Localizing document tampering is extremely challenging, as manipulations are crafted to appear visually consistent and often leave only subtle traces that are nearly invisible to the human eye. In prior work, evaluation has been largely dominated by synthetic benchmarks that closely match the training distribution, and methods have shown steady progress under this setting. However, these gains often translate poorly to human-made forgeries and to cross-domain evaluation, where both the source documents and the tampering pipeline can change, leading to a distribution shift. In addition, since the introduction of the Frequency Perception Head for the discrete cosine transform (DCT) modality, it has become a standard choice, and subsequent work has largely focused on downstream modules and fusion strategies rather than revisiting the backbone itself. To help close this gap in cross-domain performance and improve the DCT backbone design, we propose \textbf{DiffNet}, a relatively simple yet effective RGB--DCT early-fusion architecture driven by two key design choices. First, to ensure that the decoder aggregates multi-scale inconsistency evidence rather than operating on raw, content-heavy activations, we apply a lightweight multi-level discrepancy transformation at the output of each backbone stage, replacing features with magnitude-only responses to learned zero-sum filters. Second, we design an efficient DCT-domain backbone that relies on a lightweight frequency-index-aware DCT--quantization joint embedding. Our approach achieves state-of-the-art performance on cross-domain and human-made document tampering localization, outperforming prior methods by around 30\%, with up to $7\times$ higher throughput than the previous best model.

Problem

• document tampering localization
• cross-domain evaluation
• distribution shift
• DCT backbone
• human-made forgeries

Innovation

• multi-level discrepancy features
• DCT-quantization embedding
• cross-domain tampering localization
• early-fusion architecture
• frequency-index-aware embedding