Earthvision Virtual Workshop together with the CVPR Conference have set out two challenges for the event´s edition of this year. The DynamicEarthNet Challenge proposed in conjunction with DLR, TUM and Planet aims to stimulate innovation in spatio-temporal machine learning to improve the development of models without the necessity for large-scale annotations, which is often not available in practice. It will award the winners with some fellowships of the Beyond Fellows program of the AI4EO Future Lab at the Technical University of Munich.
The challenge is centered around modeling multi-temporal land cover changes (with few or no training labels) from Planetscope and Sentinel time series data.
Challenge Start: March 1, 2021
Challenge Deadline: June 1, 2021
Take a look here for further details, send your proposal and join our scientific team!
