Bento Packaging Activity Recognition Challenge

Submission Dates:
06/01/2021 to 08/01/2021
Citation Author(s):
Sayeda Shamma
Alia
Kyushu Institute of Technology
Kohei
Adachi
Kyushu Institute of Technology
Nazmun
Nahid
Kyushu Institute of Technology
Haru
Kaneko
Kyushu Institute of Technology
Paula
Lago
Kyushu Institute of Technology
Sozo
Inoue
Kyushu Institute of Technology
Submitted by:
Sayeda Shamma Alia
Last updated:
Mon, 06/07/2021 - 00:44
DOI:
10.21227/cwhs-t440
Data Format:
Links:
License:
Creative Commons Attribution

Abstract 

Human activity recognition (HAR) has been one of the most prevailing and persuasive research topics in different fields for the past few decades. The main idea is to comprehend individuals’ regular activities by looking at bits of knowledge accumulated from people and their encompassing living environments based on sensor observations. HAR has a great impact on human-robot collaborative work, especially in industrial works. In compliance with this idea, we have organized this year’s Bento Packaging Activity Recognition Challenge. In this challenge, participants will have to design methods to identify different activities done during Bento-box packaging from MOCAP data. The data has been collected during the Bento-box packaging sessions inside the lab. We will organize a workshop at ABC Conference to discuss the pros and cons of the different approaches shared by the participants of the challenge and future perspectives of this topic.

Instructions: 

Rules

The Bento Packaging Activity Recognition Challenge is organized by Sozolab.

Registration

Participants should accept all the terms and conditions in this page before registering. By registering, participants agree to these terms. Participants should also register as workshop participants in ABC. Participants not registered in the conference will not be considered in the final leaderboard of the challenge nor for the Prizes.

Elegibility

Challenge is open to students, graduate students, researchers, professors, and data scientists. Members of Sozolab (“Organizers”) are not eligible to enter or win. Only submissions with a submitted paper are eligible to win a prize.

Workshop Participation

All participants are encouraged to participate in the Workshop to share details about their approach.

Data use

All participants may use the data free of charge under CC license.

Team Limits

The maximum size of a team is 10 participants.

Competition Dataset Files

AttachmentSize
File Train_data_bento.zip959.71 MB
ACCESS ON AWS