Informal skill education transcending time and space

The objective of this research is to develop a method that automatically segments, labels, and summarizes instructional records from situations where skilled individuals instruct novices, utilizing speech and action patterns during effective skill learning. The aim is to enable cross-temporal and cross-spatial skill transmission. First-person perspective videos captured by head-mounted cameras worn by workers will be segmented, labeled, and used to generate instructional content on a per-task basis. Segmentation and labeling remain challenging tasks, but we address them by focusing on the characteristics of informal learning.