geoclass-image - GEOCLASS-image v2.0

With the release of GEOCLASS-image (v2.0) comes general improvements for ease of use. Some of these additions include visualization and manipulation of training datasets without the use of the graphical user interface (GUI), visualization of variograms while image labeling and more user customizability through the configuration folder. The larger changes to the GEOCLASS-image cyberinfrastructure (CI) revolve around the versatility of inputs, approaches, and outputs available within the CI.

The most influential changes revolve around the addition of a prototype NN VarioNet. VarioNet combines VarioMLP and ResNet-18 to make classifications based on structural and visual features. Once classifications are made, the CI can now be used to save these labeled datasets. The user has several options while saving the datasets which allow for the creation of large scale labeled datasets that can be useful for the future machine learning applications.

- Python
Published by Herzfeld-Lab about 1 year ago

geoclass-image - GEOCLASS-image v1.0

Initial release of the GEOCLASS-image software.

- Python
Published by Herzfeld-Lab about 2 years ago