NBA Multimodal Video Analysis
Title: NBA Multimodal Video Analysis
DNr: Berzelius-2024-207
Project Type: LiU Berzelius
Principal Investigator: Karol Wojtulewicz <>
Affiliation: Linköpings universitet
Duration: 2024-06-28 – 2025-01-01
Classification: 10207


Multi-modal processing of NBA videos for caption generation using transformers network. In the project various pre-trainig and finetuning methods will be used. Projects final goal is a scientific paper within the area