This project is a part of Data Science hobby projects of mine.
The purpose of this project is to use machine learning and computer vision techniques to convert the Black and White Videos to color. Thanks for Deoldify model created by Jason Antic. DeOldify is based on the fast.ai library — which brings more power and optimization for deep learning developers.DeOldify is using Generative Adversarial Networks with the iterative interplay between two Neural Networks Generator and Discriminator.
- Machine Learning
- Video GAN
- Deoldify
- Python
- Jupyter
I have fed short videos of old black and white from youtube and Imgur to run the model and generate the colored version of the videos. Some of the results can be found below
- Clone this repo (for help see this tutorial).
- You can find the videos or gif's from sources like Youtube or imgur
- I suggest you to run the notebook in Colab if you don't have a personal GPU and dont forget to set the Runtype to 'GPU'
- The process time is propotional to the length and size of the videos. I suggest first try with gif's as takes relatively less time for process.
- The processes videos can be found in Deoldify --> results folder
If interested you can checkout some of the videos generated by me using this model by clicking on the links below.
Please find the notebook to get a quick grasp at the whole project.
Fast.AI - https://github.com/fastai
Name | Git Hub Handle | Website |
---|---|---|
[Tejesh Batapati] | https://github.com/tejeshb | https://tejesh-ai.webflow.io/ |
Feel free to look at my personal projects website here [tejesh.AI] (https://tejesh-ai.webflow.io/)