The purpose of the project is to implement a semantic image-to-image translation using Generative Adversarial Networks and Autoencoders to elucidate on unsupervised machine learning. The intent is to develop a GANs based model to create new novel images from existing image input.

Create Autoencoder model to condense feature selection on Image dataset.
Formulate a GAN to perform image to Image translation.
Create a Pix2Pix Model to train and deploy the conditional generative adversarial network.

The target audience for this Deep Learning project is a user who wants to design new shoes designs from the earlier existing data .Other uses of the model are listed below:
To provide help to designers for designing Provide users or their clients with diverse ideas.
To create new images form one reference.
The uses and possibilities are endless. Use of the model depends on the user’s perspective and creative intellect.

Flutter
Dart
Python
Flask(backend)

Google Colab
VS Code
Android Studio

Edges2shoes Dataset

Youtube