The 19th International Conference on Document Analysis and Recognition (ICDAR 2025), organized by the School of Software Engineering at Huazhong University of Science and Technology (HUST), was held in Wuhan from Sept 16 to 21, attracting 431 scholars and industry representatives from 31 countries and regions.

ICDAR 2025 featured keynote speeches, tutorials, a doctoral forum, and multiple thematic workshops, showcasing the latest developments in document analysis and recognition (DAR) in the era of large models and multimodal AI.

Professor Koichi Kise from Osaka Metropolitan University shared personal insights in his speech “From AI to AI – Why Document Analysis and Recognition Stands Out”, highlighting DAR’s unique role in artificial intelligence.

Professor Qiao Yu from Shanghai Artificial Intelligence Laboratory discussed advances and challenges in large multimodal models for document understanding.

Professor Josep Lladós from the Computer Vision Center at Universitat Autonoma de Barcelona introduced the “dual syntax” framework for documents, encompassing both visible structures, such as text and layout, and hidden structures that span paragraphs, pages, or entire documents.

A roundtable discussion on “Beyond Foundation Models: The Next Challenges in Document Analysis” further explored future directions in the field.

Professor Bai Xiang, Dean of HUST’s School of Software Engineering, served as one of the ICDAR 2025 chairs. Next year, the conference will be held in Vienna, Austria.

Source: School of Software Engineering of HUST