# Towards Zero-Shot Differential Morphing Attack Detection with Multimodal Large Language Models

arXiv: https://arxiv.org/abs/2505.15332 · HTML: https://mlsanity.com/paper/2505.15332/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in MLLM jailbreaks and adversarial robustness, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 11 of its 45 references that are in the map.
- First posted on arXiv 21 May 2025.

## Placed in

- Multimodal LLMs / vision-language models / [MLLM jailbreaks and adversarial robustness](https://mlsanity.com/mllm/mllm-jailbreaks-and-adversarial-robustness/)

## Builds on (most cited first, 11 of 11)

- [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://mlsanity.com/paper/2201.11903/) (2022)
- [Face Morphing Attack Generation & Detection: A Comprehensive Survey](https://mlsanity.com/paper/2011.02045/) (2020)
- [Deep Face Representations for Differential Morphing Attack Detection](https://mlsanity.com/paper/2001.01202/) (2020)
- [Attention Aware Wavelet-based Detection of Morphed Face Images](https://mlsanity.com/paper/2106.15686/) (2021)
- [MADation: Face Morphing Attack Detection with Foundation Models](https://mlsanity.com/paper/2501.03800/) (2025)
- [Generalized Single-Image-Based Morphing Attack Detection Using Deep Representations from Vision Transformer](https://mlsanity.com/paper/2501.09817/) (2025)
- [Differential Morphed Face Detection Using Deep Siamese Networks](https://mlsanity.com/paper/2012.01541/) (2020)
- [Algorithmic Fairness in Face Morphing Attack Detection](https://mlsanity.com/paper/2111.12115/) (2021)
- [Reliable Face Morphing Attack Detection in On-The-Fly Border Control Scenario with Variation in Image Resolution and Capture Distance](https://mlsanity.com/paper/2209.15474/) (2022)
- [MorphGANFormer: Transformer-based Face Morphing and De-Morphing](https://mlsanity.com/paper/2302.09404/) (2023)
- [Differential Newborn Face Morphing Attack Detection using Wavelet Scatter Network](https://mlsanity.com/paper/2305.01294/) (2023)

## Built on by (most cited first, 2 of 2)

- [Benchmarking Foundation Models for Zero-Shot Biometric Tasks](https://mlsanity.com/paper/2505.24214/) (2025)
- [DifFoundMAD: Foundation Models meet Differential Morphing Attack Detection](https://mlsanity.com/paper/2604.17961/) (2026)

## Abstract (arXiv)

Leveraging the power of multimodal large language models (LLMs) offers a promising approach to enhancing the accuracy and interpretability of morphing attack detection (MAD), especially in real-world biometric applications. This work introduces the use of LLMs for differential morphing attack detection (D-MAD). To the best of our knowledge, this is the first study to employ multimodal LLMs to D-MAD using real biometric data. To effectively utilize these models, we design Chain-of-Thought (CoT)-based prompts to reduce failure-to-answer rates and enhance the reasoning behind decisions. Our contributions include: (1) the first application of multimodal LLMs for D-MAD using real data subjects, (2) CoT-based prompt engineering to improve response reliability and explainability, (3) comprehensive qualitative and quantitative benchmarking of LLM performance using data from 54 individuals captured in passport enrollment scenarios, and (4) comparative analysis of two multimodal LLMs: ChatGPT-4o and Gemini providing insights into their morphing attack detection accuracy and decision transparency. Experimental results show that ChatGPT-4o outperforms Gemini in detection accuracy, especially against GAN-based morphs, though both models struggle under challenging conditions. While Gemini offers more consistent explanations, ChatGPT-4o is more resilient but prone to a higher failure-to-answer rate.