That's a VIBE: ML for Human Pose and Shape Estimation with Nikos Athanasiou, Muhammed Kocabas, Michael Black - #409

This article from Practical AI discusses the research paper "VIBE: Video Inference for Human Body Pose and Shape Estimation" submitted to CVPR 2020. The podcast episode features Nikos Athanasiou, Muhammed Kocabas, and Michael Black, exploring their work on human pose and shape estimation using an adversarial learning framework. The conversation covers the problem they are addressing, the datasets they are utilizing (AMASS), the innovations distinguishing their work, and the experimental results. The article provides a brief overview of the research, highlighting key aspects like the methodology and the datasets used, and points to the full show notes for more details.

• •The research focuses on human pose and shape estimation from video.
• •The study utilizes an adversarial learning framework.
• •The work builds upon existing datasets like AMASS.