################################################ https://gdo152.llnl.gov/cowc/download/ Parent Directory cowc-everything.tar This is everything in the COWC directory compressed into one file. cowc-m-everything.tar This is everything in the COWC-M directory compressed into one file. cowc-m/ This is the new COWC-M dataset which has finer grained labels for the type of car. cowc/ This is the original COWC dataset described in our ECCV 16 paper. ################################################ https://gdo152.llnl.gov/cowc/download/cowc Parent Directory DISCLAIMER.txt LICENSE.txt datasets/ This directory contains the COWC dataset and related image sets. networks/ This directory contains pre-trained networks in Caffe format. papers/ This contains the original paper describing COWC published in ECCV 2016. readme.txt result_images/ These are result images that show counting and detection on the COWC final testing set. ################################################ https://gdo152.llnl.gov/cowc/download/cowc/datasets/ Parent Directory IM-813368 ground_truth_sets/ These are large overhead images which have been annotated with ground truth data. The original annotations are provided in Gimp image format. However, positive and negative images are in png format. patch_sets/ These are images that have already been extracted and are ready for use in training/testing. readme.txt testing_scenes/ These are the final held-out testing scenes. These images are large format and are the same images used to generate the ones seen in https://gdo152.llnl.gov/cowc/download/cowc/result_images/ ################################################ https://gdo152.llnl.gov/cowc/download/cowc/datasets/ground_truth_sets/ Each of the original ground truth sets is in its own directory sorted by the original dataset it was derived from. ################################################ https://gdo152.llnl.gov/cowc/download/cowc/datasets/patch_sets/ counting/ The patch training set used for counting training. detection/ The patch training set used for detection training. ################################################ https://gdo152.llnl.gov/cowc/download/cowc-m/ Parent Directory datasets/ This is the COWC-M dataset. How to use the contents is described here: https://github.com/LLNL/cowc/tree/master/COWC-M networks/ This is an empty directory