I am currently pursuing the M.S. degree in information and communication engineering at University of Electronic Science and Technology of China, supervised by Prof. Guolong Cui. Before that, I obtained my B.S. degree from School of Information Science & Engineering at Lanzhou University, supervised by Prof. Kun Zhan. My research focuses on radar target detection and tracking via deep learning. CV / Google Scholar / GitHub Last updated July 2023 |
Radar targets are generally non-cooperative, resulting in a significant scarcity of target samples. Even if there are a small amount of valuable data, accurately labeling them is challenging since the true position of the target is unknown. However, most data-driven detection methods focus on training with a accurate label set.
The goal of my research is to make object recognition available in the real world. The models should be trained without precise labels.
Towards this goal, my research focuses on three aspects: 1. How to extract meaningful representations from raw data. I propose a method for radar visual representations via contrastive learning and its application on target detection (RAVA-CL). 2. How to learn from incomplete or imprecise annotations. I propose a weakly supervised radar target detection method to detect potential targets and generate range cell-level labels (WSRTD). 3. How to generalize object recognition through time. I propose an end-to-end tracking framework based on learning displacement of the target in successive frames. (TBLD).
Going further, I am interested in target detection, target tracking, supervised learning, and unsupervised learning.