A5062092763- Advanced Vision and Imaging
- Robotics and Sensor-Based Localization
- 3D Surveying and Cultural Heritage
- 3D Modeling in Geospatial Applications
- Computer Graphics and Visualization Techniques
- Radiation Dose and Imaging
- Remote Sensing and LiDAR Applications
- Optical measurement and interference techniques
- Advanced X-ray and CT Imaging
- Advanced Radiotherapy Techniques
Group Sense (China)
2021-2023
Sensetime (China)
2021
Southern Medical University
2016
We present a real-time monocular 3D reconstruction system on mobile phone, called Mobile3DRecon. Using an embedded camera, our provides online mesh generation capability back end together with 6DoF pose tracking front for users to achieve realistic AR effects and interactions phones. Unlike most existing state-of-the-art systems which produce only point cloud based models or surface offline, we propose novel incremental approach fast dense satisfy the demand of applications. For each...
Urban scene modeling is a challenging task for the photogrammetry and computer vision community due to its large scale, structural complexity, topological delicacy. This paper presents an efficient multistep framework large-scale urban scenes from aerial images. It takes images textured 3D mesh model generated by image-based system as input outputs compact polygon models with semantics at different levels of detail (LODs). Based on key observation that buildings usually have piecewise planar...
This paper presents an accurate and robust dense 3D reconstruction system for detail preserving surface modeling of large-scale scenes from multi-view images, which we named DP-MVS. Our performs high-quality reconstruction, preserves geometric details thin structures, especially linear objects. framework begins with a sparse carried out by incremental Structure-from-Motion. Based on the reconstructed map, novel PatchMatch approach is applied depth estimation each image view. The estimated...
Consumer-level RGB-D cameras have been widely used for dense 3D reconstruction of scenes. Especially textureless or non-lambertian surfaces, consumer can ensure completeness the reconstructed models at a low cost. However, quality relies heavily on accuracy depth sensors. Digital are also popularly capturing high-resolution pictures to achieve high-quality scenes, but cannot handle regions well due visual ambiguity problem. To both and models, we propose hybrid multi-view pipeline named...
The purpose of this study is to design and fabricate an anthropopathic abdominal phantom for accuracy evaluation deformable image registration (DIR) algorithms in adaptive radiation therapy.The constructed organs, including the liver, kidney, spleen stomach, are made mixture polyvinyl chloride (PVC) softener dioctyl terephthalate, while rigid structures, i.e. vertebrae, white cement.Relation between PVC-softener blending ratio organ CT number studied, three-dimensional printing technic...
This paper presents an accurate and robust Structure-from-Motion (SfM) pipeline named LiVisSfM, which is SfM-based reconstruction system that fully combines LiDAR visual cues. Unlike most existing LiDAR-inertial odometry (LIO) LiDAR-inertial-visual (LIVO) methods relying heavily on registration coupled with Inertial Measurement Unit (IMU), we propose a LiDAR-visual SfM method innovatively carries out frame to voxel map in Point-to-Gaussian residual metrics, combined BA explicit loop closure...