DQRNet: Dynamic Quality Refinement Network for 3D Reconstruction from a Single Depth View DOI Creative Commons
Caixia Liu,

Minhong Zhu,

Haisheng Li

et al.

Sensors, Journal Year: 2025, Volume and Issue: 25(5), P. 1503 - 1503

Published: Feb. 28, 2025

With the widespread adoption of 3D scanning technology, depth view-driven reconstruction has become crucial for applications such as SLAM, virtual reality, and autonomous vehicles. However, due to effects self-occlusion environmental occlusion, obtaining complete error-free shapes directly from scans remains challenging, previous methods tend lose details. To this end, we propose Dynamic Quality Refinement Network (DQRNet) reconstructing accurate shape a single view. DQRNet introduces dynamic encoder–decoder detail quality refiner generate high-resolution shapes, where former designs latent extractor adaptively select important parts an object latter global local point refiners enhance quality. Experimental results show that is able focus on capturing details at boundaries key areas ShapeNet dataset, thereby achieving better accuracy robustness than SOTA methods.

Language: Английский

DQRNet: Dynamic Quality Refinement Network for 3D Reconstruction from a Single Depth View DOI Creative Commons
Caixia Liu,

Minhong Zhu,

Haisheng Li

et al.

Sensors, Journal Year: 2025, Volume and Issue: 25(5), P. 1503 - 1503

Published: Feb. 28, 2025

With the widespread adoption of 3D scanning technology, depth view-driven reconstruction has become crucial for applications such as SLAM, virtual reality, and autonomous vehicles. However, due to effects self-occlusion environmental occlusion, obtaining complete error-free shapes directly from scans remains challenging, previous methods tend lose details. To this end, we propose Dynamic Quality Refinement Network (DQRNet) reconstructing accurate shape a single view. DQRNet introduces dynamic encoder–decoder detail quality refiner generate high-resolution shapes, where former designs latent extractor adaptively select important parts an object latter global local point refiners enhance quality. Experimental results show that is able focus on capturing details at boundaries key areas ShapeNet dataset, thereby achieving better accuracy robustness than SOTA methods.

Language: Английский

Citations

0