Statoil, an international energy company needs to identify drifting icebergs in North Atlantic Ocean for safe navigation. They use a computer vision model built by C-CORE from satellite data to classify objects either as a ship or a drifting iceberg.
This code uses deep learning computer vision models on the satellite data provided by Statoil/C-Core to identify icebergs. Specifically, it uses Inception-ResNet-V2 network. Final prediction was an ensemble of solutions provided by Inception-ResNet-V2, ResNet50, Inception-V3, and VGG16 networks.