With the rapid development of AIGC technology, detecting AI-generated fake images has become a new research hotspot. However, existing detection methods vary in experimental setups and datasets, making it difficult to directly compare their performance. To address this, we conduct a comprehensive analysis and comparison of mainstream AIGC image detection methods and propose a new approach with improved performance. Furthermore, we provide an integrated benchmarking platform for testing and evaluation. Under consistent training datasets and experimental conditions, the platform evaluates detection accuracy, generalization, and other metrics. It serves as a standardized benchmark for AIGC image detection research. We also integrate and open-source various existing detection algorithms for the research community. We welcome you to download the source code, conduct your own evaluations, and cite our arXiv paper.
MSAQE is an innovative project developed at Fudan University that leverages AI to assess the quality of tourist attractions based on visitor reviews. By analyzing over 6 million comments, the system provides accurate real-time evaluations to support informed decision-making for government users.
Fudan MAS Lab