Description and Areas of Research
My research interests are in computer vision and machine learning. My recent research focuses on deep learning theories and optimization, and applications of deep learning to non-Euclidean data.
Research in Geometric Perception and Intelligence Research (Gorilla) lab is featured by two types of awareness of data geometry: machine learning models that exploit inherent geometric regularities of high-dimensional sensory data, and perception methods that are tailored for data with non-Euclidean geometric structures.
We expect such research can potentially bring innovations in numerous applications ranging from cloud based visual semantics to AI empowered robots. The ultimate goal of Gorilla lab is to develop practical AI solutions that can naturally sense, analyze, and interact with the visual world.
- Deep Learning and Optimization with Robustness and Stability (under construction)
- Learning with Less Supervision via Domain Adaptation (under construction)
- Deep Learning Shape Modeling and Reconstruction of Objects and Scenes (under construction)
- Semantic 3D Analysis with Full Degrees of Freedom (under construction)
- 3D Object Detection and Segmentation in Videos, sponsored by Huawei Technologies, 2020 - 2021
- 3D Content Creation for Learning and from Learning, sponsored by Microsoft Research Asia, 2020
- Deep Learning Multi-view Object Surface Reconstruction, sponsored by Huawei Technologies, 2019 - 2021
- TOF based High-performance 3D Sensing for Vision-guided Autonomous and Intelligent System, sponsored by The Guangdong Key R&D Program, 2020 - 2022
- Theoretical Foundation, Key technologies, and Applications of Intelligent Mega-Media Perception, sponsored by The Program for Guangdong Introducing Innovative and Enterpreneurial Team, 2018 - 2023
- Optimization of Deep Neural Networks and the Geometric and Physical Solution Properties, sponsored by NSFC of China, 2018 - 2021
Others Relevant to Research