CAVS Semantic Segmentation Dataset for Off-Road Autonomous Driving (CaSSeD)
***************************************************************************************************
Authors: Suvash Sharma, Lalitha Dabbiru, Chris Goodin, Daniel Carruth, Matthew Doude, Christopher Hudson, John Ball, Bo Tang
***************************************************************************************************

The package named "CaSSeD_Dataset.tar.gz" consists of real world and virtual datasets. It consists of raw images and the corresponding annotations along with rosbags. 

It can be extracted in linux with command: tar -xvf CaSSeD.tar.gz
In windows it can be extracted simply with any unzipping software.

The "real-world" folder contains rosbags from where the training and testing images were extracted. This offroad dataset contains 1679 real-world RGB camera images and the corresponding annotations. The virtual offroad dataset folder "MAVS" contains 1947 images and annotations. 
The "Train" folder contains 350 images and corresponding annotations. The "annos" folder for annotations contains another folder "int_maps" where the integer mapped pixel values are assigned for each classes. 
The "Test" folder contains 150 test images altogether inside "mixed" folder. For three different tracks individually, 60 images are in "Browns_field”, 60 images in "Main_trail", and 30 images in "Powerline" folder. int_maps are also provided for each test set.
Dataset2 contains fog data with 279 images, Dataset 3 contains 300 images and Dataset 4 contains 600 images.
 ***************************************************************************************************
If you use this dataset for research, please consider citing our dataset paper entitled "CaSSeD: CAVS Semantic Segmentation Dataset for Off-Road Autonomous Driving".

***************************************************************************************************