Henghui Ding

Papers

Total Citations

261

H-Index

About

Henghui Ding is a leading researcher in computer vision, specializing in visual segmentation, open-vocabulary object tracking, and transformer-based architectures. His most impactful contribution is the comprehensive survey "Transformer-Based Visual Segmentation" (2024), which has garnered 192 citations and serves as a definitive resource for partitioning images, video frames, and point clouds—critical for autonomous driving, medical analysis, and robotics. Ding also pioneered OVTrack (2023), an open-vocabulary multiple object tracking system that extends traditional MOT beyond a few categories to recognize and track countless dynamic objects in real-world scenes, earning 58 citations. His work bridges the gap between closed-set benchmarks and the open-world demands of self-driving and robotic systems. With a focus on deep learning innovations, Ding’s research has significantly advanced the field’s ability to handle diverse, unconstrained environments. His surveys and methods are widely cited by practitioners and academics alike, cementing his reputation as a key figure in modern visual understanding and a go-to source for students and engineers tackling complex segmentation and tracking challenges.