作者: Daniel Gongora , Hikaru Nagano , Masashi Konyo , Satoshi Tadokoro
DOI: 10.1007/978-3-319-93399-3_54
关键词: Artificial intelligence 、 Vibration 、 Terrain 、 Transient (computer programming) 、 Computer science 、 First person 、 Computer vision 、 Collision detection
摘要: Fast playback of First-Person View (FPV) videos reduces watching time but it also increases the perceived intensity camera trembling and makes transient events, such as collisions, less evident. Here we propose using vibrations vibrotactile feedback to support collision detection in fast video playback. To preserve pitch during playback, use Time-Scale Modification (TSM) methods developed for audio. We show that delivered palm dominant hand improved performance a pilot study. found reducing levels terrain is beneficial detection. Furthermore, without participants are likely underestimate number collisions video. Our results suggest has potential events FPV videos.