Ziheng Ding

Papers

Total Citations

H-Index

About

Ziheng Ding is a rising researcher in 3D computer vision, autonomous driving, and robotics, whose work focuses on advancing perception and localization systems. His major contributions include pioneering adaptive sampling strategies for 3D object detection and developing robust LiDAR-visual SLAM frameworks. In his paper "AS-Det: Active Sampling for Adaptive 3D Object Detection in Point Clouds" (2025, 2 citations), Ding introduced a novel active sampling mechanism that overcomes the limitations of traditional point-based detectors, enabling more effective local representation learning from unstructured raw point clouds—a critical step for real-world autonomous systems. He further advanced the field with "DeepPointMap2: Accurate and Robust LiDAR-Visual SLAM with Neural Descriptors" (2024, 1 citation), where he replaced hand-crafted feature extraction with learned neural descriptors, significantly improving cross-modal fusion and robustness in simultaneous localization and mapping. While still early in his career, Ding's work demonstrates a clear trajectory toward solving fundamental challenges in 3D perception, with potential to impact autonomous driving and robotics applications. His research is particularly notable for addressing the gap between simplistic sampling methods and the complex demands of real-world point cloud processing.