https://github.com/mpolinowski/tf-mnist-digits
Tensorflow unsupervised learning to denoise images from the mnist digits dataset
https://github.com/mpolinowski/tf-mnist-digits
denoise-images mnist-image-dataset representation-learning tensorflow2 unsupervised-learning
Last synced: about 1 month ago
JSON representation
Tensorflow unsupervised learning to denoise images from the mnist digits dataset
- Host: GitHub
- URL: https://github.com/mpolinowski/tf-mnist-digits
- Owner: mpolinowski
- Created: 2022-12-21T05:20:55.000Z (over 2 years ago)
- Default Branch: master
- Last Pushed: 2022-12-21T05:20:59.000Z (over 2 years ago)
- Last Synced: 2025-01-28T19:17:28.900Z (3 months ago)
- Topics: denoise-images, mnist-image-dataset, representation-learning, tensorflow2, unsupervised-learning
- Language: Python
- Homepage: https://mpolinowski.github.io/docs/IoT-and-Machine-Learning/ML/2022-12-20-tf-representation/2022-12-19
- Size: 51.8 KB
- Stars: 0
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Tensorflow Representation Learning
Use Tensorflow/Keras to build an __Autoencoder__ to perform an unsupervised learning to de-noise image files. The Autoencoder is based on a regular Keras sequential model and consists of the encoder - that is basically the same convolution layer used in regular CNNs with supervised learning - and an decoder - that is an "encoder in reverse". The decoder takes the compressed representation of the image generated by the encoder from the representation layer and has to reconstruct the initial image from this compress data:
```py
# build autoencoder model
autoencoder = tf.keras.models.Sequential()
# build the encoder CNN
autoencoder.add(tf.keras.layers.Conv2D(16, (3,3), strides=1, padding="same", input_shape=(28, 28, 1)))
autoencoder.add(tf.keras.layers.MaxPooling2D((2,2), padding="same"))
autoencoder.add(tf.keras.layers.Conv2D(8, (3,3), strides=1, padding="same"))
autoencoder.add(tf.keras.layers.MaxPooling2D((2,2), padding="same"))
# representation layer
autoencoder.add(tf.keras.layers.Conv2D(8, (3,3), strides=1, padding="same"))
# build the decoder CNN
autoencoder.add(tf.keras.layers.UpSampling2D((2, 2)))
autoencoder.add(tf.keras.layers.Conv2DTranspose(8,(3,3), strides=1, padding="same"))
autoencoder.add(tf.keras.layers.UpSampling2D((2, 2)))
autoencoder.add(tf.keras.layers.Conv2DTranspose(1, (3,3), strides=1, activation='sigmoid', padding="same"))
```
For the training we now feed the a noisy dataset into the encoder and have the Autencoder compress and reconstruct each image. After those steps we provide to original - noise-free - image to compare the reconstructed image to. The training is complete when we reach a minimum of differences between both of them:
```py
# fit model to dataset
autoencoder.fit(X_train_noisy.reshape(-1, 28, 28, 1),
X_train.reshape(-1, 28, 28, 1),
epochs=10,
batch_size=200)
```
As an example, here are 15 "noised" images from the mnist digits dataset and the reconstructed de-noised image the Autoencoder generated from it:
