Skip to content

A detailed study of state-of-the-art Image Colorization ML models: Deoldify, Instance-Aware, and ChromaGAN

Notifications You must be signed in to change notification settings

guptaviha/ML-Image-Colorization-Study

Repository files navigation

Image Colorization: A Comparitive Study

An academic project for Intro To Machine Learning (CS-GY-6923) at NYU Tandon with Prof Fraida Fund.

Objective: To study a state-of-the-art method in machine learning by: learning about your method, replicating some existing version of it (using a Python notebook or source code that is provided for you), and then extending it.

For more details: Prompt

Colorization of grayscale images is an ill-posed problem, with multiple correct solutions. Although important progress has been made in this field, over time multiple models have been implemented to solve this problem. We compare three popular models in detail:

  1. Deoldify
  2. Instance-Aware Image Colorization
  3. ChromaGAN

Test Images can be found here on GitHub. Code and implementation details are available in our Colab Notebook. Open In Colab

Team

Disha Lamba - dl4747@nyu.edu

Viha Gupta - vg2237@nyu.edu

About

A detailed study of state-of-the-art Image Colorization ML models: Deoldify, Instance-Aware, and ChromaGAN

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published