Coast Train is a dataset made using Doodler, and has been used for training Zoo models that are being implemented in CoastSeg, Seg2Map, and elsewhere.Provides a graphical browser-based environment for application of Segmentation Zoo models for mapping features in satellite imagery.Provides a graphical browser-based environment for application of the CoastSat workflow for shoreline mapping.A mapping extension for CoastSat using Segmentation Zoo models. Uses codes housed in s2mengine for conversion of label images (from Gym and Doodler) into geospatial formats, and for visualization of generic label images in geospatial formats in a small webGIS viewer.A mapping extension for application of Segmentation Zoo models on geospatial imagery.Dash-Doodler and Holo-Doodler both use the same codes for image label generation, found in doodler_engine repo.A port of Dash-Doodler using the Anaconda/Holoviews API for graphical user interface.Packages are compatible, and share an underlying design and data structures □ Applications and datasets built on top of Doodler, Gym, and Zoo Holo-Doodler You are encouraged to contribute your Gym models for the common good!.A set of notebooks that illustrate how best to use a model on sample imagery.A repository of pre-trained Gym models for some common tasks.This repository is where you train models based on your own imagery and label data, and evaluate those models.A neural gym for training image segmentation models based on fully convolutional models based on UNets.
0 Comments
Leave a Reply. |