Jinghua Liu
Papers
2
Total Citations
6
H-Index
2
About
Jinghua Liu is a researcher advancing the field of 3D computer vision, with a primary focus on 6D object pose estimation—a critical technology for robotics, augmented reality, and autonomous systems. Liu’s work centers on learning robust representations from monocular images, addressing the fundamental challenge of inferring an object’s full 3D position and orientation from a single 2D view. In their highly cited 2024 paper, "Spatial and temporal consistency learning for monocular 6D pose estimation," Liu introduced a novel framework that leverages both spatial geometry and temporal coherence across video frames to achieve more accurate and stable pose predictions, garnering 4 citations. Building on this, their 2025 study, "Images, normal maps and point clouds fusion decoder for 6D pose estimation," pioneered a multi-modal fusion decoder that integrates RGB images, surface normal maps, and 3D point clouds, achieving 2 citations and demonstrating a powerful approach to combining complementary data sources. These contributions are notable for pushing the boundaries of monocular pose estimation, making it more reliable in real-world, dynamic environments. Liu’s work is essential reading for students and researchers seeking to understand how consistency learning and sensor fusion can overcome the inherent ambiguities of single-image 3D reasoning.
Research Focus
Key Achievements
Top Papers
- 1
- 2