The doodleverse is an opinionated collection of Python packages designed for geoscientific image segmentation.
Package maintainers:
- @dbuscombe-usgs Marda Science / USGS Pacific Coastal and Marine Science Center. Developed originally for the USGS Coastal Marine Geology program, as part of the Florence Supplemental project
- @ebgoldstein
Currently, there are 3 packages:
- Dash-Doodler
- For fast, interactive segmentation of imagery
- For creation of training datasets for training Machine-Learning-based image segmentation models, for example Segmentation Gym
- Segmentation Gym
- A neural gym for training image segmentation models based on fully convolutional models based on UNets
- Segmentation Zoo
- A repository of pre-trained models
Packages are compatible, and share an underlying design and data structures