A5054117340- 3D Shape Modeling and Analysis
- Remote Sensing and LiDAR Applications
- Advanced Neural Network Applications
- Robotics and Sensor-Based Localization
- Advanced Vision and Imaging
- Leaf Properties and Growth Measurement
- Human Pose and Action Recognition
- Smart Agriculture and AI
McGill University
2023
Reliable and automated 3-dimensional (3D) plant shoot segmentation is a core prerequisite for the extraction of phenotypic traits at organ level. Combining deep learning point clouds can provide effective ways to address challenge. However, fully supervised methods require datasets be point-wise annotated, which extremely expensive time-consuming. In our work, we proposed novel weakly framework, Eff-3DPSeg, 3D segmentation. First, high-resolution soybean were reconstructed using low-cost...
Semantic understanding of 3D point clouds is important for various robotics applications. Given that point-wise semantic annotation expensive, in this paper, we address the challenge learning models with extremely sparse labels. The core problem how to leverage numerous unlabeled points. To end, propose a self-supervised representation framework named viewpoint bottleneck. It optimizes mutual-information based objective, which applied on under different viewpoints. A principled analysis...