Score: 0

Detecting Text Manipulation in Images using Vision Language Models

Published: September 12, 2025 | arXiv ID: 2509.10278v1

By: Vidit Vidit , Pavel Korshunov , Amir Mohammadi and more

Potential Business Impact:

Finds fake writing in pictures better.

Business Areas:
Image Recognition Data and Analytics, Software

Recent works have shown the effectiveness of Large Vision Language Models (VLMs or LVLMs) in image manipulation detection. However, text manipulation detection is largely missing in these studies. We bridge this knowledge gap by analyzing closed- and open-source VLMs on different text manipulation datasets. Our results suggest that open-source models are getting closer, but still behind closed-source ones like GPT- 4o. Additionally, we benchmark image manipulation detection-specific VLMs for text manipulation detection and show that they suffer from the generalization problem. We benchmark VLMs for manipulations done on in-the-wild scene texts and on fantasy ID cards, where the latter mimic a challenging real-world misuse.

Page Count
12 pages

Category
Computer Science:
CV and Pattern Recognition