Graph learning, sampling and filtering for image and signal estimation
Graph signal processing (GSP) is the study of computational tools for correlated data residing on irregular kernels described abstractly by graphs. I overview recent progress in GSP theories and their applications to image and more general signal estimation tasks. I first overview approaches in learning appropriate graph structures given limited observed data. I illustrate how good graphs can lead to competitive and robust graph-based classification results in a semi-supervised learning setting. I then discuss fast graph sampling schemes--generalizing Nyquist sampling in traditional signal processing to the graph domain--and their applications to practical engineering scenarios, such as matrix completion and 3D point cloud sub-sampling. Finally, I discuss how analytical graph filters can be combined with deep learning tools in a hybrid framework for robust, explainable and memory-efficient implementations of image denoising algorithms.
Date and Time
Location
Hosts
Registration
-
Add Event to Calendar
- Starts 09 March 2021 03:20 AM UTC
- Ends 29 March 2021 07:59 PM UTC
- No Admission Charge
Speakers
Prof. Gene Cheung of York University
Graph learning, sampling and filtering for image and signal estimation
Graph signal processing (GSP) is the study of computational tools for correlated data residing on irregular kernels described abstractly by graphs. I overview recent progress in GSP theories and their applications to image and more general signal estimation tasks. I first overview approaches in learning appropriate graph structures given limited observed data. I illustrate how good graphs can lead to competitive and robust graph-based classification results in a semi-supervised learning setting. I then discuss fast graph sampling schemes--generalizing Nyquist sampling in traditional signal processing to the graph domain--and their applications to practical engineering scenarios, such as matrix completion and 3D point cloud sub-sampling. Finally, I discuss how analytical graph filters can be combined with deep learning tools in a hybrid framework for robust, explainable and memory-efficient implementations of image denoising algorithms.
Biography:
Gene received his Ph.D. degree in electrical engineering and computer science from the University of California, Berkeley in 2000. He was a senior researcher in Hewlett-Packard Laboratories Japan, Tokyo, from 2000 till 2009. He was an assistant then associate professor in National Institute of Informatics (NII) in Tokyo, Japan, from 2009 till 2018. He is now an associate professor in York University, Toronto, Canada.
His research interests include 3D imaging and graph signal processing. He has served as associate editor for multiple journals, including IEEE Transactions on Multimedia (2007--2011), IEEE Transactions on Circuits and Systems for Video Technology (2016--2017) and IEEE Transactions on Image Processing (2015--2019). He served as a member of the Multimedia Signal Processing Technical Committee (MMSP-TC) in IEEE Signal Processing Society (2012--2014), and a member of the Image, Video, and Multidimensional Signal Processing Technical Committee (IVMSP-TC) (2015--2017, 2018--2020). He is a co-author of several paper awards, including the best student paper award in ICIP 2013, ICIP 2017 and IVMSP 2016, best paper runner-up award in ICME 2012, and IEEE Signal Processing Society (SPS) Japan best paper award 2016. He is a recipient of the Canadian NSERC Discovery Accelerator Supplement (DAS) 2019. He is a Fellow of IEEE.
Address:4700 Keele St., , Toronto, Ontario, Canada, M3J 1P3