DOI: 10.18178/wcse.2025.06.016
Automatic Gaming Highlights Generation Using Facial Expression Analysis of Game Streamer
Abstract— With the spread of short video platforms, there has been a growing demand to generate short highlights from very long video live streams especially gaming live streams, which has become a burden for video editors. Therefore, it is necessary to introduce an automatic highlights generation method to reduce the burden on editors, while providing viewers with an efficient viewing experience. Existing works on gaming highlights generation relied on viewer comments, and thus cannot analyze videos with few comments. Moreover, since the comments are view-driven information, highlights generated based on comments may not correctly reflect the intention of the streamer. This study focuses on information obtained from the video itself. Specifically, the emotional responses of game streamers based on their facial expressions are utilized for automatically extracting highlight scenes from long gaming live streams.
Index Terms— Video Summarization, Facial Expression, Emotion Analysis
Kazuma Kii, Eiji Kamioka, Chanh Minh Tran, Phan Xuan Tan
Graduate School of Engineering and Science, Shibaura Institute of Technology, JAPAN
Cite: Kazuma Kii, Eiji Kamioka, Chanh Minh Tran, Phan Xuan Tan, "Automatic Gaming Highlights Generation Using Facial Expression Analysis of Game Streamer", 2025 the 15th International Workshop on Computer Science and Engineering (WCSE 2025), pp. 97-101, Jeju Island, South Korea, June 28-30, 2025.