Spot the Fake: Large Multimodal Model-Based Synthetic Image Detection with Artifact Explanation

TL;DR

The paper introduces FakeVLM, a specialized large multimodal model designed for both detecting synthetic images and generating natural-language artifact explanations, addressing interpretability gaps in existing detectors. To train and evaluate it, the authors create FakeClue, a 100k-sample, seven-category dataset annotated for fine-grained artifacts through a multi-LMM pipeline and category priors. Empirical results across FakeClue, LOKI, and FF++ demonstrate that FakeVLM matches or surpasses expert detectors on authenticity tasks while providing high-quality artifact explanations, all without additional classifiers. The work highlights the potential of category-aware prompts, multi-model annotation, and QA-style explanations to achieve transparent, robust synthetic image detection with practical impact for verifying visual content.

Abstract

With the rapid advancement of Artificial Intelligence Generated Content (AIGC) technologies, synthetic images have become increasingly prevalent in everyday life, posing new challenges for authenticity assessment and detection. Despite the effectiveness of existing methods in evaluating image authenticity and locating forgeries, these approaches often lack human interpretability and do not fully address the growing complexity of synthetic data. To tackle these challenges, we introduce FakeVLM, a specialized large multimodal model designed for both general synthetic image and DeepFake detection tasks. FakeVLM not only excels in distinguishing real from fake images but also provides clear, natural language explanations for image artifacts, enhancing interpretability. Additionally, we present FakeClue, a comprehensive dataset containing over 100,000 images across seven categories, annotated with fine-grained artifact clues in natural language. FakeVLM demonstrates performance comparable to expert models while eliminating the need for additional classifiers, making it a robust solution for synthetic data detection. Extensive evaluations across multiple datasets confirm the superiority of FakeVLM in both authenticity classification and artifact explanation tasks, setting a new benchmark for synthetic image detection. The code, model weights, and dataset can be found here: https://github.com/opendatalab/FakeVLM.

Figures (8)

• Figure 1: Construction pipeline of FakeClue dataset, including data collection from open source and self-synthesized datasets, pre-processing with categorization, label prompt design based on category knowledge(Face M: Face manipulation), and multiple LMMs annotation with result aggregation.
• Figure 2: Overview of FakeVLM, our proposed framework for detecting synthetic images and explaining their artifacts. Built upon LLaVA, FakeVLM integrates multiple captioning models to assess key visual aspects.
• Figure 3: Comparison of synthetic image detection approaches on LOKI and FakeClue datasets: (1) QA with Frozen LMMs (no training), (2) Frozen backbone + linear probe (only linear layer trained), (3) Direct Real/Fake QA tuning, and (4) VQA with artifact explanations tuning.
• Figure 4: Synthetic image detection cases, covering animals, people, objects, documents, and remote sensing (red denotes incorrect, green denotes correct detection). FakeVLM outperforms GPT in precision, comprehensiveness, and relevance, demonstrating superior detection and interpretation.
• Figure 5: Performance of FakeVLM on real images.