Andrew Zhai

Very high resolution canopy height maps from RGB imagery using self-supervised vision transformer and convolutional decoder trained on aerial lidar

Billion-Scale Pretraining with Vision Transformers for Multi-Task Visual Representations