Python scripts to perform semantic segmentation of driving images generated with the CARLA simulator, making use of a pytorch implementation of DeepLabv3. The notebooks included are:
-
generate_semantic_segmentation_data.ipynb
: CARLA python script to generate semantic segmentation dataset, creating RGB images and semantic masks. -
semantic_segmentation_carla.ipynb
: train DeepLabv3 model for semantic segmentation in CARLA images.
Sample data included in CARLA_smalldataset.tar.xz
.