Computer Vision News - May 2024

Computer Vision News 2 A Great New Paper! Sai Kumar Dwivedi (left) is a PhD student at the Max Planck Institute for Intelligent Systems MPI-IS, supervised by MPI-IS Director Michael J. Black and Dimitrios Tzionas (right), an assistant professor at the University of Amsterdam and formerly of MPI-IS, with which he continues to collaborate. Sai and Dimitrios speak to us about their novel framework, POCO, which extends human pose and shape regressors to output both 3D bodies and uncertainty together. Conventionally, human pose and shape regressors have focused solely on predicting a body’s pose and shape without considering the confidence level associated with the output. However, for downstream applications, understanding the uncertainty of these predictions is crucial to ascertaining their reliability and usefulness. POCO: 3D Pose and Shape Estimation using Confidence