Predicted segmentation masks from UNet, Dense UNet and DeepLabv3+ for breech and firing-pin areas for a NIST dataset of firearm cartridge case images

The dataset in 10.15132/10000262 contains training, validation and test data. The test data consist of 339 images. We have used this data to train three CNNS: UNet, Dense UNet and DeepLabv3+. We have calculated the predicted masks for the test data at 512x512 pixel resolution for the firing-pin, breech face areas as well and both areas together. A post-processing step was perform on the masks. The dataset contains the mask from the three CNNs, the three areas and with and without post-processing.

These data are made available under Creative Commons Attribution Licence CC-BY 4.0 International https://creativecommons.org/licenses/by/4.0/deed.en