Team MSU GenText-Forensics Challenge 2026 Technical Report

Kirill Koltsov; Aleksandr Gushchin; Dmitriy Vatolin; Anastasia Antsiferova

cs.CVSep 29, 2026

Team MSU GenText-Forensics Challenge 2026 Technical Report

Authors: Kirill Koltsov, Aleksandr Gushchin, Dmitriy Vatolin, Anastasia Antsiferova

Organizations: Lomonosov Moscow State University Moscow, Russia · MSU Institute for Artificial Intelligence Moscow, Russia

Abstract

Document text forgery has evolved beyond simple pixel-level manipulation: modern attacks alter not only the appearance of a document but also its meaning, and increasingly target the OCR & LLM pipelines that consume such documents. The ACM MM 2026 GenText-Forensics challenge therefore requires systems that not only decide whether a multilingual text image is forged, but also localize the point of manipulation, identify the attack type, and produce a human-readable forensic report with supporting evidence. We present our solution, a decomposed chain-of-thought (CoT) pipeline that combines a document tampering detector (DTD) with two Qwen3-VL-32B vision-language models, each LoRA-adapted to a distinct sub-task. DTD produces tampering probability maps that are converted into numbered candidate regions; a first model (the Filterer) validates these regions and assigns a preliminary forgery type, while a second model (the Semantic Detective) merges and re-grounds the surviving regions, searches for purely semantic anomalies that are invisible to pixel-level detectors, and writes the final report. Both models are trained by distilling chain-of-thought traces from a privileged Qwen3-VL-235B teacher that has access to ground-truth masks and reports. Our approach secured third place in the ACM MM 2026 GenText-Forensics challenge. We describe the data preparation, test-time augmentation, region rendering, distillation protocol, and training configuration in detail, and report ablations over detector thresholds, prompt designs, and pipeline decompositions.

Figures & tables

Figure 1. Overview of the decomposed chain-of-thought pipeline. Document icons labeled "txt" denote the natural-language prompts supplied to the Qwen models. These prompts are not fixed: a distinct template is used at each stage, and each template is populated dynamically with stage-specific evidence before generation. Prior to the Qwen Filterer, the prompt is augmented with the numbered DTD candidate regions (each index matching a red box on the annotated image). Prior to the Qwen Semantic Detective, the prompt is augmented with only the regions the Filterer marked KEEP, together with the full output of the OCR model—the recognized text of every word and its bounding-box coordinates in the source document—so that the model can ground its findings on observed token positions. Block diagram of the five pipeline stages from input document image to final forensic report.

Filtering of DTD regions	Final Score
Clean image	$0.659$
Clean Image + heatmap overlay	$0.751$
Image with numbered red boxes	$\mathbf{0.794}$

Table 1. Choice of visual–textual input, measured on all 500 held-out test images with Two-Stage pipeline

Filtering of DTD regions	mIoU	mF1
None (raw DTD regions)	$0.6402$	$0.7058$
ResNet / U-Net filter	$0.6113$	$0.6510$
Qwen Filterer (ours)	$\mathbf{0.6672}$	$\mathbf{0.7404}$

Table 2. Effect of the Qwen Filterer on downstream localization, measured on all

500

held-out test images (forged and pristine).

Setting	Value
Base model	Qwen3-VL-32B-Instruct
Quantization	4-bit NF4, double quant., bf16 compute
LoRA rank $r$	$16$
LoRA $\alpha$	$32$
LoRA dropout	$0.05$
Target modules	q,k,v,o_proj

Table 3. QLoRA fine-tuning configuration for the two students.

Configuration	Trained	$S_{\mathrm{Det}}$	$S_{\mathrm{Loc}}$	$S_{\mathrm{Exp}}$	$S_{\mathrm{Rep}}$	Final
strong-direct	No	$\mathbf{1.00}$	$0.55$	$0.86$	$61.2$	$0.764$
direct	No	$\mathbf{1.00}$	$0.55$	$0.86$	$61.6$	$0.768$
all-in-one	No	$0.82$	$0.22$	$0.86$	$51.6$	$0.566$
two-stage	No	$0.92$	$0.51$	$0.86$	$56.8$	$0.721$
strong-direct	Yes	$0.97$	$0.62$	$\mathbf{0.88}$	$68.4$	$0.798$
direct	Yes	$0.95$	$0.48$	$\mathbf{0.88}$	$64.1$	$0.748$

Table 4. Comparison of prompt / pipeline configurations on the

N{=}100

-document evaluation set. “Trained” indicates LoRA fine-tuning on the corresponding targets, as opposed to zero-shot prompting of the base model.

S_{\mathrm{Det}}

S_{\mathrm{Loc}}

and

S_{\mathrm{Exp}}

lie in

[0,1]

; the LLM-judge column (

S_{\mathrm{Rep}}

) is on a

0

–

100

scale. The final submission is in bold.

Explore similar work

SEED: Simple ViT and Evolving Harness for Explainable Text Forgery Detection
Jun 19, 2026Kahim Wong, Kemou Li, Yiming Chen +2Face Forgery Detection Digital Forensics
State Key Laboratory of Internet of Things for Smart City, Department of Computer and Information Science, University of Macau · School of Information and Software Engineering, University of Electronic Science and Technology of China
ForensicsTok: Forensics-Guided Tokenized Modeling for Image Tampering Localization
Jun 23, 2026Lei Xu, Haowei Wang, Shen Chen +3Mask2Former Digital Forensics
1Shenzhen University · 2Tencent Youtu Lab
Trustworthy Image Authentication using Forensic Knowledge Graphs
Jun 22, 2026Tai D. Nguyen, Matthew C. StammDigital Forensics Knowledge Graphs
Drexel University, Philadelphia, PA 19104, USA

Jun 19, 2026, cs.CVJ/K move · Enter open · S save

SEED: Simple ViT and Evolving Harness for Explainable Text Forgery Detection

Kahim Wong, Kemou Li, Yiming Chen, Haiwei Wu+1

State Key Laboratory of Internet of Things for Smart City, Department of Computer and Information Science, University of Macau · School of Information and Software Engineering, University of Electronic Science and Technology of China

Face Forgery Detection Digital Forensics Text Analysis and Detection Self-Supervised Vision Transformers Dinov3 Evolving Harness

Open paper AskPDF arXiv

AI-assisted image editing threatens trust in financial, legal, and identity records. The GenText-Forensics Challenge at ACM MM 2026 addresses this by requiring structured forensic reports, in which integrating detection, pixel-level localization, and natural language explanation for multilingual text-centric forgery images. We present SEED, a modular system with three components. First, a similarity-guided pipeline augments training with diverse synthetic forgeries. Second, a single ViT, built on DINOv3 with LoRA adaptation, jointly performs detection and pixel-level localization while preserving pre-trained priors with minimal trainable parameters. Third, an evolving harness takes the detector's predictions and generates a complete forensic report via an MLLM, iteratively improved through a proposer-evaluator loop optimizing report quality. SEED ranked 3rd in the GenText-Forensics Challenge. Code and data are available at https://github.com/KahimWong/GenText-Forensics-3rd-Place.

Abstract

Explore similar work

SEED: Simple ViT and Evolving Harness for Explainable Text Forgery Detection

ForensicsTok: Forensics-Guided Tokenized Modeling for Image Tampering Localization

Trustworthy Image Authentication using Forensic Knowledge Graphs