Apologies for cross-posting!!
We are glad to announce the 2nd Edition of MUSIA (Multilingual Story Illustration) Track at FIRE 2026.
Track Title: Multilingual Story Illustration: Bridging Cultures through AI Artistry (MUSIA)
Venue: Forum for Information Retrieval Evaluation (FIRE 2026)
Track Website: https://cse-iitbhu.github.io/irlab/#/musia
Problem Statement:
The MUSIA 2026 track focuses on culturally grounded multimodal story understanding and generation across multiple languages, including English, Hindi, Bengali, and Marathi. The track addresses key challenges in generating and reasoning over visual narratives while preserving narrative coherence, character identity, cultural context, and artistic consistency.
Subtask 1: Story-to-Image Sequence Generation
Participants are required to generate a coherent sequence of illustrations from a given story while maintaining narrative flow and visual consistency.
Subtask 2: Anchor-Conditioned Sequence Completion
Given a story and an anchor illustration, participants must generate the remaining illustrations while preserving character identity, style, and narrative progression.
Subtask 3: Narrative Image Ordering
Participants are required to reconstruct the correct narrative sequence from a set of shuffled story illustrations.
For more details and registration, please visit: https://cse-iitbhu.github.io/irlab/#/musia
Organisers: