Since 2012, Chiba Institute of Technology Planetary Exploration Research Center (hereinafter PERC) has been working on the project called "Project Meteor", where long-term continuous meteor observations using ultra-high-sensitivity CMOS color high-vision cameras from International Space Station (ISS) are made. Actually, it's quite difficult to manually detect the meteor occurrence in the captured video as the duration of meteor illumination is usually less than 1 sec and it's impossible to predict when and where it occurs.
Therefore, STAIR Lab presents you a challenge to develop an algorithm which automatically detects the meteor in the video. The competition is organized by Software Technology and Artificial Intelligence Research Laboratory (hereinafter STAIR Lab) and uses data from PERC.
Task Description
An algorithm should take 3 sec motion video as input and classify it into two groups based on the meteor occurrence. Train set contains 10000 videos whilst test set does 1000 and the number of videos with the meteor occurrence in the train set is 200. The video format is mpg and its resolution is 320 x 180.
High school student department
We've additionally set up a department for high school student.
The task and rules are the same as what are applied to general participants.
Please see the FAQ page for details.
Contact Information
For any questions and inquiries, please contact the following address:
STAIR Lab, Chiba Institute of Technology
Email: challenge2018@stair.center