Explore a 20-minute conference talk on the discovery of latent 3D keypoints through end-to-end geometric reasoning. Delve into the KeypointNet framework, including its goals, setup, and architecture. Learn about multi-view consistency loss, relative pose estimation loss, and the importance of keypoints. Examine quantitative and qualitative results, including failure cases and ablation studies. Understand how this semi-supervised approach combines keypoint and geometry learning networks, outperforming supervised methods. Gain insights into additional testing and proof-of-concept applications for this innovative technique in 3D computer vision.
Discovery of Latent 3D Keypoints via End-to-End Geometric Reasoning
University of Central Florida via YouTube
Overview
Syllabus
Discovery of Latent 3 Keypoints via End-to- Geometric Reasonin
Overview
Problem
KeypointNet: The Goal (Testing)
KeypointNet: The Setup (Training) Image
Multi-view Consistency Loss
Relative Pose Estimation Loss
Regarding Keypoints (p. 2)
Keypoint Net: Architecture
Testing Methodology
Quantitative Results (p. 2)
Qualitative Results (p. 2)
Qualitative Results p. 3, failure ca
Additional Results (ablation, primary losses)
Additional Results (other testing)
Additional Results proof-of-concept Imag
More Information
Summary • Semi-supervised end-to-end keypoint find • Combines keypoint and geometry learning network • Outperforms supervised method
Taught by
UCF CRCV
