U-Inception-Net: semantic segmentation of aerial images for land use/cover mapping

This project is an research for my Data Science Specialization Thesis at ITA. The article for publication is in progress...

We propose a new Deep Learning architecture for semanatic segmentation of aerial image combining the U-Net with the Inception module.

• The core folder contains all the code in python with classes to use the models.
• The Notebooks folder contains some Notebooks with the training and test of the models.
• The Fake Dataset folder contains some code to generate an fake dataset to train and test the models.