IEEE

Self-Supervised Learning for Motion Synthesis and Prediction in Animated Characters

Abstract: Self-Supervised Learning for Motion Synthesis and Prediction in Animated Characters uses a new system that uses motion capture data that hasn't been marked to learn detailed models of how ...
IEEE

SharpDepth: Sharpening Metric Depth Predictions Using Diffusion Distillation

Abstract: We propose SharpDepth, a novel approach to monocular metric depth estimation that combines the metric accuracy of discriminative depth estimation methods (e.g., Metric3D, UniDepth) with the ...