https://github.com/tanyachutani/denoising-convolution-auto-encoder

Keras implementation of denoising convolution auto encoder on mnist
https://github.com/tanyachutani/denoising-convolution-auto-encoder

autoencoder cnn deep-learning denoising-autoencoders keras keras-neural-networks mnist tensorflow

Last synced: 3 months ago
JSON representation

Keras implementation of denoising convolution auto encoder on mnist

• Host: GitHub
• URL: https://github.com/tanyachutani/denoising-convolution-auto-encoder
• Owner: TanyaChutani
• Created: 2020-04-05T16:15:15.000Z (over 6 years ago)
• Default Branch: master
• Last Pushed: 2020-04-12T14:09:50.000Z (about 6 years ago)
• Last Synced: 2025-02-27T17:21:58.755Z (over 1 year ago)
• Topics: autoencoder, cnn, deep-learning, denoising-autoencoders, keras, keras-neural-networks, mnist, tensorflow
• Language: Jupyter Notebook
• Homepage:
• Size: 51.8 KB
• Stars: 0
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# Denoising_ConvolutionAutoEncoder

Denoising Convolution AutoEncoder on mnist

## Data 

Mnist dataset from keras.datasets

## Model

- Autoencoder, a deep neural network is implemented with Keras functional API.

- Input to model is noisy image and output is the same image without noise.

- Filter size is 16 followed by 32 and latent dimesion is 16 units.

- The loss function for training is mse.

## Result

You can download the pre-trained weights from here 


https://drive.google.com/open?id=1KmnNJM29HADuNUIIM8MxNmDfkM9wTDgN

#### Input 


![](https://github.com/TanyaChutani/Denoising_ConvolutionAutoEncoder/blob/master/noisyA.png)


#### Output 


![](https://github.com/TanyaChutani/Denoising_ConvolutionAutoEncoder/blob/master/outputA.png)