Association football (soccer) is one of the most popular sports worldwide, and makes up a considerable share of the global streaming traffic. However, the production of distribution-ready multimedia content for soccer often requires expensive equipment and manual labor. Automating the video production pipeline and providing fast game highlights at a much lower cost is therefore desired, and developments in Artificial Intelligence (AI) have shown great potential in this context. This talk introduces a number of enabling technologies, proof-of-concept implementations, and professional deployment examples, based on deep learning for efficiently automating soccer multimedia production. Target use cases include automatic event detection, highlight clipping, thumbnail selection, game summarization, and storytelling on social media, with the common goal of almost-real-time multimedia production (audio, video, text, and combinations of modalities) from VoD/live streams. Experiences and lessons learned from both research and industry are also disseminated.

Presented by: Cise Midoglu / Simula Research Library Streaming Tech Sweden: https://streamingtech.se