Shypd.aiThe largest open index of AI toolsEPro-PnP is an AI Agents & Automation tool that provides a generalized end-to-end probabilistic Perspective-n-Points (PnP) solution for monocular object pose estimation. It was recognized as the Best Student Paper at CVPR 2022.
About
EPro-PnP is a cutting-edge probabilistic Perspective-n-Points (PnP) layer designed for end-to-end 6DoF pose estimation networks. This tool, recognized as the CVPR 2022 Best Student Paper, offers a differentiable approach to the PnP operation, which is traditionally non-differentiable. By outputting a posterior distribution of pose approximated via Monte Carlo sampling, EPro-PnP enables learning correspondences from scratch. It supports both 6DoF pose estimation (EPro-PnP-6DoF) and 3D object detection (EPro-PnP-Det), with a V2 model achieving state-of-the-art performance on benchmarks like nuScenes for camera-based single-frame object detection. The project provides models and a demo for integration into other models.
Capabilities
Pricing & Plans
Open Source
Free
FAQs
