HANYANG UNIVERSITY CNA Laboratory
Deep Learning on 3D Point Clouds
* Core problems on 3D geometric data such as point clouds include semantic segmentation, object detection and instance segmentation. The primary obstacle is that point clouds are inherently unordered, unstructured and non-uniform.
* Point Cloud 3d object detection compute object features from the irregularly and sparsely distributed points while Instance segmentation processes the point clouds to output a category and an instance mask for each detected object. These two serving as the foundation for a wide range of applications such as autonomous driving, virtual reality, and robot navigation.
Semi-supervised Object Detection
* Despite the great progress in both 2D and 3D object detection, most works focused on a fully-supervised setting. A few works have proposed to leverage unlabeled data or weakly-annotated data for 2D object detection. Aim at detecting the amodal oriented 3D bounding boxes of object along with the semantic class label. In particular, accomplishing this task under challenging conditions of limited supervision where only have access to a small set of labeled scenes and a set of unlabeled scenes.