# cnn-audio-denoiser

**Repository Path**: mathematicsX/cnn-audio-denoiser

## Basic Information

- **Project Name**: cnn-audio-denoiser
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2021-05-28
- **Last Updated**: 2021-11-03

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# A Fully Convolutional Neural Network for Speech Enhancement

Tensorflow 2.0 implementation of the paper [A Fully Convolutional Neural Network for Speech Enhancement](https://pdfs.semanticscholar.org/9ed8/e2f6c338f4e0d1ab0d8e6ab8b836ea66ae95.pdf)

Blog post: [Practical Deep Learning Audio Denoising](https://medium.com/better-programming/practical-deep-learning-audio-denoising-79c1c1aea299)

## Dataset

Part of the dataset used to train the original system is now [available to download](http://cdn.daitan.com/dataset.zip)
  
The zip file contains 1 training file (that is 10% of the data used to train the system), a validation file, and two 
audio files (not included in the training files) used to evaluate the model. 

You can create the dataset for yourself. 

- Download the [Mozilla Common Voice](https://voice.mozilla.org/) and [UrbanSound8K](https://urbansounddataset.weebly.com/urbansound8k.html) datasets.
- Use the ```create_dataset.py``` script to create the TFRecord files.