Panoptic Segmentation of Satellite Image Time Series with Convolutional Temporal Attention Networks

Benchmark Model Rank Results
cloud-removal-on-sen12ms-cr-tsU-TAE#2RMSE: 0.051PSNR: 27.05SSIM: 0.849SAM: 11.649
flood-extent-forecasting-on-global-floodU-TAE#1F1 score: 0.77
panoptic-segmentation-on-pastisU-TAE + PaPs#3PQ: 40.4RQ: 49.2SQ: 81.3
semantic-segmentation-on-pastisU-TAE#3Mean IoU (test): 63.1Number of Params: 1.1MOverall Accuracy: 83.2