EVENTA: Event-Enriched Image Analysis Grand Challenge 

The EVENTA Challenge at ACM Multimedia 2025 focuses on advancing image-caption retrieval and generation, with a goal to enrich image captions with contextual details related to events. 

The challenge includes two main tasks:

• Event-Enriched Image Captioning – Participants generate captions that go beyond simple visual descriptions by offering richer insights, including information about objects, context, and event significance.

• Event-Based Image Retrieval – Given a caption, participants retrieve the corresponding image from a database, improving search efficiency and context-awareness in retrieval systems.

The challenge will release a dataset containing over 50,000 image-caption pairs curated from credible articles. The challenge emphasizes the integration of event-related context to generate richer image descriptions. Evaluation will be based on metrics like BLEU, ROUGE, CIDEr, and more for captioning, and precision and recall for retrieval.

• Trung-Nghia Le (University of Science, VNU-HCM, Vietnam) - Email: ltnghia@fit.hcmus.edu.vn

• Tam V. Nguyen (University of Dayton, USA) - Email: tamnguyen@udayton.edu

• Khanh-Duy Le (University of Science, VNU-HCM, Vietnam) - Email: lkduy@fit.hcmus.edu.vn

• Phuc-Tan Nguyen (University of Science, VNU-HCM, Vietnam)

• Trong-Hieu Nguyen (University of Science, VNU-HCM, Vietnam)

• Quang Nguyen (University of Science, VNU-HCM, Vietnam)

• Thien-Phuc Tran (University of Science, VNU-HCM, Vietnam)

• Minh-Triet Tran (University of Science, VNU-HCM, Vietnam) - Email: tmtriet@fit.hcmus.edu.vn

• TBA