(De)Noise



Mentors :

  • Omkar Ghugarkar

  • Bhuvan Aggarwal

Mentees :

  • 10


Imagine waiting for your flight at the airport. Suddenly, an important call lights up your phone. Tons of background noise clutters up the soundscape around you — background chatter, airplanes taking off, maybe a flight announcement. You have to take the call and you want to sound clear. We all have been in this awkward, non-ideal situation. It’s just part of modern business. Background noise is everywhere. And it’s annoying.

In this project, we would do Speech Enhancement. Speech enhancement is the task of taking a noisy speech input and producing an enhanced speech output.
How are we going to do?
We would be using focusing on deep learning techniques. Initially we will use autoencoders then move onto implement SEGAN (1703.09452)
Selection Procedure
Your selection would be based on your SOP and a small coding test in Python (If you are good with CS101 and have an intro to python, it should be a cake-walk)
Caution
This project would require good amount of commitment. But believe me, you would enjoy the journey and learn lots of Cool stuff.
Perks
Introduction to the whole new world of machine Learning and Deep learning Master the Deep learning packages like Tensor Flow, keras Apply the hottest technique in Deep Learning – GANs Get acquainted to various techniques in speech processing Last but not the least, apply your coding skills to build an entire Project from scratch
Resources
Python Tutorial - https://www.youtube.com/playlist?list=PL-osiE80TeTskrapNbzXhwoFUiLCjGgY7
Speech Enhancement - https://medium.com/analytics-vidhya/noise-suppression-using-deep-learning-6ead8c8a1839
Prerequisites: A good understanding of concepts of CS – 101 is always good. It would be preferable if you have an introduction to Python. Don’t worry if you are not very comfortable with Python. Please go through the small Python tutorial attached in the resources section. Along with this, interest to learn new things and enthusiasm is must.

Tentative Timeline :

Week Work
Week 1 Introduction to Linear Regression and Logistic regression
Week 2 Get acquainted with Neural Networks and Deep learning
Week 3 Get acquainted with CNN's and RNN's and Resnet architecture
Week 4 Implement the autoencoder model
Week 5-6 Read, understand and start implementing the GAN model of the Project using Tensor-Flow
Week 7 Buffer