Recognizing Activities of Daily Living (ADL) is a vital process forintelligent assistive robots . But collecting large annotated datasets require temporal labeling and raises privacy concerns . We build Sims4Action by specifically executingactions-of-interest in a “top-down” manner . We integrate two modern algorithms for video-based activityrecognition in our framework . We will make ourdataset publicly available at https://://://github.com/aroitberg/sims4action. We will also make our dataset public available athttps://://www.geneitberg.com/. The results also indicate that tasks involving a mixture of gaming andreal data are challenging, opening a new research direction. However, our results also show that tasks involve a mixture and real data is challenging, however, may be challenging, open a newResearch direction. We believe that task involving a combination of games and data are very difficult to use. We

Author(s) : Alina Roitberg, David Schneider, Aulia Djamal, Constantin Seibold, Simon ReiƟ, Rainer Stiefelhagen

Links : PDF - Abstract

Code :
Coursera

Keywords : action - data - challenging - sims - video -

Leave a Reply

Your email address will not be published. Required fields are marked *