Jianfang Chang
Papers
2
Total Citations
17
H-Index
2
About
Jianfang Chang is a researcher in robotics and computer vision, whose work focuses on advancing visual simultaneous localization and mapping (VSLAM) systems—the technology that enables autonomous robots and drones to navigate unknown environments. Chang’s key contributions lie in improving loop closure detection, a critical component that allows a robot to recognize when it has returned to a previously visited location, thereby correcting cumulative drift in its map. To tackle this challenge, Chang has pioneered the use of deep metric learning, specifically through triplet loss functions. Their 2021 paper introduced a triplet-loss-based approach for closed-loop detection, and their 2022 follow-up, "Weighted triplet loss based on deep neural networks for loop closure detection in VSLAM," refined this method by incorporating weighted sampling strategies. These works, each garnering 9 and 8 citations respectively, have been influential in demonstrating how neural networks can learn more robust and discriminative visual features compared to traditional handcrafted descriptors. By enhancing the reliability of VSLAM systems, Chang’s research directly supports safer and more efficient autonomous navigation in applications ranging from warehouse logistics to search-and-rescue missions.
Research Focus
Key Achievements
Top Papers
- 1
- 2