RhizoVision Transformer

RhizoVision Transformer is planned as a community resource by which the community shares raw root imagery without compromising their own work or data ownership, a vision transformer is trained on Frontier at ORNL, and the pre-trained model is shared as an open weight and open source. The community can then use these models to develop various 'heads' or decodes for specific tasks such as segmentation or beyond.

RhizoVision Explorer will be updated to support Windows and Linux, a command line version for scripting, and updates to trait measurements.

Please take a moment to answer a few questions in this short survey to show interesting in sharing imagery or providing feedback for the development of RhizoVision.

RhizoVision Survey