2019 EE331 - Colorization

Organized by MinjeongJu - Current server time: June 19, 2019, 2:58 p.m. UTC

First phase

First phase
May 27, 2019, midnight UTC

End

Competition Ends
June 18, 2019, midnight UTC

2019 - EE331 Colorization

This is the project for 2019 EE331 Introduction to Machine Learning in KAIST.

When you participate this competition, please let me know your student IDs. If not, you will not be approved.

 

Our goal is to make colorization from input images.

The normal images are composed of RGB values, but we can represent these images into LAB values.

In here, L means lightness.

We cannot know the color of the image without the A and B values.

 

If we know only the L value of the image, let's say that the image is L image.

Our detail task is as follows:

    1. Given LAB train images, we have to construct the model and train it.

        - The input of the model should be L images only with lightness.

        - The output of the model can be A, and B values or LAB images (You can choose what you want)

    2. Given L test images, we have to reconstruct the LAB images using the trained model.

 

 The ppt file from Team2, and the explanation of cost and score are contained below link

            https://drive.google.com/drive/folders/1zG_sb-8oGsFwin4E63_ns-0Xsv5A0_kB?usp=sharing

The detail explanation of data and submission format are described in 'Data' page of 'Participate' menu.

First phase

Start: May 27, 2019, midnight

Competition Ends

June 18, 2019, midnight

You must be logged in to participate in competitions.

Sign In