Hi! I am a Global Research Assistant Professor (GRAP Scheme) from the College of Computing (Data Science) at City University of Hong Kong. Before that, I was a research fellow and a Ph.D. student in the Faculty of Engineering at the National University of Singapore (NUS) in NUS supervised by Prof. Chee Yeow Meng. I received an M. Sc (Electrical & Computer Engineering) degree from NUS. I was also hired as a research assistant at IORA, NUS, working on several innovative projects of deep learning visual applications. Besides the vision-based projects, I am also interested in implementing machine learning techniques in medical applications. My research work was published in distinguished conferences (three oral presentation papers) and journals, including CVPR, ICCV, NeurIPS, ACM MM, IJCAI, AAAI, IJCV, IEEE TITS and IEEE JSTSP. I am seeking motivated students or postdoctoral researchers to join my research team at CityU. Interested candidates are encouraged to send an email to lixkedu@outlook.com.

My Current Research Topics:

• Deep Learning on 3D Geometry
• Trustworthy AI
• Deep Learning for Healthcare

News

• May. 3, 2025: A journal paper “Learning Against Distributional Uncertainty: On the Trade-off Between Robustness and Specificity” was accepted by IEEE Journal of Selected Topics in Signal Processing (JSTSP).
• Apr. 30, 2025: A corresponding-authored paper “Enhancing Sampling Protocol for Point Cloud Classification Against Corruptions” was accepted by IJCAI 2025
• Mar. 7, 2025: A corresponding-authored journal paper “Deep Hierarchical Learning for 3D Semantic Segmentation” was accepted by the International Journal of Computer Vision (IJCV).
• Jul. 28, 2024: A co-first-authored paper “Pyramid Diffusion for Fine 3D Large Scene Generation” was accepted and selected as Oral Prepsentation in ECCV 2024. The implementation code was released in [GitHub link](https://github.com/shinke-li/pointcvar.
• Feb. 25, 2024: Accepted paper “PointCVaR: Risk-optimized Outlier Cleanse for Robust Point Cloud Classification” was selected as Oral Prepsentation in AAAI 2024. The implementation code was released in GitHub link.
• Dec. 09, 2023: A first-authored paper “PointCVaR: Risk-optimized Outlier Cleanse for Robust Point Cloud Classification” was accepted by AAAI 2024.
• Dec. 02, 2023: A co-first-authored paper “A Railway Accident Prevention System Using An Intelligent Pilot Vehicle” was accepted by IEEE TITS.
• Jun, 23, 2022: I am glad to attend the offline meeting at New Orleans, CVPR 2022.
• Apr. 16, 2022: A first-authored paper “Primitive3D: 3D Object Dataset Synthesis From Randomly Assembled Primitives” was accepted by CVPR 2022
• Sep. 28, 2021: A paper “Directed Graph Contrastive Learning” was accepted by NeurIPS 2021.
• Jul. 22, 2021: A first-authored paper “PointBA: Towards Backdoor Attacks in 3D Point Cloud” was accepted by ICCV 2021.
• Jul. 03, 2021: A reproducibility paper was accepted by ACM MM 2021.
• Sep. 26, 2020: A paper “Digraph Inception Convolutional Networks” was accepted by NeurIPS 2020.
• Aug. 06, 2020: Campus3D launches publicly. The official website is here.
• Jul. 26, 2020: A first-authored paper “Campus3D: A photogrammetry point cloud benchmark for hierarchical understanding of outdoor scene” was accepted by ACM MM 2020 Oral Persentation.
• Mar. 16, 2020: Campus3D dataset has been constructed.