SAT CNN Models

Code and models for training convolutional neural networks (CNN) and satellite image classification using SAT-4 and SAT-6 datasets (http://csc.lsu.edu/~saikat/deepsat/).

Datasets in the Collection

Thumbnail of Tools & Results
Code for convnet training and image classification, conference paper and presentation.
Code and usage instructions are also available from Vi-SEEM code repository https://code.vi-seem.eu/Risojevic/cnn-features.
The implementation is in Python and it is based on Theano (https://github.com/Theano/Theano) and the neural networks library Lasagne (https://github.com/Lasagne/Lasagne). For installing Theano and Lasagne please follow the installation instruction on the respective github pages. You will also need: numpy, scipy and sklearn
Created on May 03, 2017
CNN models trained using SAT-6 dataset.
Created on Dec 26, 2016
CNN models trained using SAT-4 dataset.
Created on Dec 21, 2016

Child Collections in the Collection

There are no child collections in this collection and not enough permission to edit this collection.

Spaces containing the Collection

Parent collections