Training data for 'Segmenteverygrain'

This folder contains 48 images of grains and the corresponding segmentation masks that form the majority of the images that were used to train the U-Net model that the 'Segmenteverygrain' Python package is based on. The images have filenames that terminate in '_image.png'; the mask filenames terminate in '_mask.png'. The mask rasters only contain three values: 0 for background, 1 for the grain itself, and 2 for the grain boundary.

These files can be used to train a new U-Net model, either using 'Segmenteverygrain' functions, or using any machine learning framework that has functionality for training image segmentation models.

Some of these images come from the SediNet project (Buscombe, 2019). A few images of fluvial gravel were collected by Mair et al. (2022), using UAVs; see this repository. The remaining images were taken either with a handheld digital camera or using a microscope.