Stagemix video generation using face and body keypoints detection

Minjoon Jung, Seunghyun Lee, Eun-Seon Sim, Min Ho Jo, Yu Jin Lee, Hyebin Choi, Junseok Kwon

Abstract: Playing multiple stage videos of a particular singer as if they are one is called Stagemix video. The consumption of video media has increased recently, and the demand for video editing has also increased. Stagemix videos have gained popularity in various communities, and a number of YouTubers who upload videos with cross-cuts are appearing. In this work, we introduce a novel task, Stagemix video generation. Stagemix video generation requires considerable time and skillful editing skills. To address this, we suggest a method of auto-generating Stagemix video, a novel technique that plays multiple stage videos of a particular singer as if they are one. Our novel methods automatically generate a Stagemix video and improve performance with face or body keypoints which is extracted by CNN-based extractor. Quantitative differences between frames and creation time show that our methods effectively produce a natural video.