Kyle Gao

PhD Candidate

Research Interests

My recent work has been in deep learning, specifically, Computer Vision and its application to point clouds and aerial orthoimages. I have recently worked on segmentation of aerial orthoimages with applications to building rooftop detection and land use/land cover classification, as well as segmentation and classification of LiDAR point clouds. Im also interested in Point Cloud Quality Assessment and Compression. I am currently investigating both geometric and deep learning-based Point Cloud Quality Assessment metrics in order find a fast Point Cloud Quality Assessment method with high correspondence to the Human Visual System.

Education

  • Master of Science, Accelerator Physics program, University of Victoria, 2017-2020
  • Honours Bachelors degree in Mathematics, Mathematical Physics Co-op program, University of Waterloo, 2011-2016

Publications

  • A region-based deep learning approach to instant segmentation of aerial orthoimagery for building rooftop detection, K. Gao, M Chen, S. N. Fatholahi, H. He, H. Xu, J. M. Junior, W. N. Gon├žalves, M. A. Chapman, J. Li. Submitted to Geomatica, Sep 2021.
  • Waterloo Building Dataset: A city-scale vector building dataset for mapping building footprints using aerial orthoimagery, H. He, Z. Jiang, K. Gao, S. N. Fatholahi, W. Tan, B. Hua, H. Xu, M. A. Chapman, J. Li. Submitted to Geomatica, July 2021.
  • Airborne Multispectral LiDAR Point Cloud Classification with a Feature Reasoning-based Graph Convolution Network, P. Zhao, D. Li, Y. Yu, H. Wang, K. Gao, H. Guan, J. M. Junior, J. Li. Submitted to the International Journal of Applied Earth Observation. July 2021.
  • Hyperspectral and Multispectral Image Fusion by Jointing Spatial-Spectral Dual-Dictionary with Structured Sparse Low-rank Representation, N. Chen, B. Zhang, H. He, K. Gao, L. Sui, J. M. Junior, J. Li. Submitted to the International Journal of Applied Earth Observation. July 2021.
  • A RandLA-Net-based Approach to Semantic Segmentation of 3D Objects in MLS Point Clouds of Large-scale Urban Roadways, X. Lu, D. Chen, Z. Wang, H. Ma, L. Cheng, M. Li, X. Ma, Y. Chen, J. M. Junior, K. Gao, J. Li. Submitted to IEEE Transaction on Intelligent Transportation Systems.
  • EGUN-ELBT reference trajectory correction in presence of ambient field, D. Kaltchev, K. Gao. TRIUMF Beam Physics Notes. May 2014.