AIC x TOO Collaboration Contest Activity Report

This is an activity report on  the AIC x Too collaboration contest (Hackathon/Ideathon) held in October 2020.

Background

In the anime, manga, and film industries, mistakes in character coloring, etc., have traditionally been detected manually. However, the automation of mistake detection using AI and similar tools is gaining traction so that creators can concentrate on creative activities. Also, the human eye tends to judge a flyer or poster’s aesthetics based on vague standards, but by creating evaluation criteria for appearance and scoring them, even new designers will be able to evaluate them objectively. With this in mind, we held a contest (Hackathon/Ideathon) sponsored by Too.

Contest Details

The theme of the hackathon was “Detecting miscolored character images in videos,” and four teams and individual participants competed to create better programs in terms of model accuracy and speed. At the presentation session on October 31, each group gave a presentation on the programs they had created, including their innovations and difficulties faced.

The theme of the Ideathon was “Creating evaluation criteria for the appearance of flyers.” Four teams, from both arts and science majors, competed in terms of usability, originality, and business models, etc. At the presentation session on October 31, each team was asked to present the evaluation criteria for the flyers that they had come up with.

At the award ceremony on November 13, Too announced the winning teams of the Hackathon and Ideathon. Each of the winning teams of the Hackathon/Ideathon received an iPad and an Insta360 GO. The winners are as follows. *Permission to publish has been obtained from each winner.

Hackathon Winners

Ideathon Winners

Mr. Koki Tomei