Center for Signal and Information Processing (CSIP) Seminar Series presents: Recursive Tensor Robust Principal Component Analysis-Mohammadpour Salut Mohammad
Date: Friday, April 2, 2021
Bluejeans link: https://bluejeans.com/621436705
Mohammad Salut received the M.S. degree in Georgia Institute of Technology in 2018. He is currently pursuing the Ph.D. degree in electrical engineering. My current research interests are in high dimensional data analysis and machine learning for multi-stream signals, images, videos, point clouds and MRI.
ABSTRACT: Online robust PCA algorithms are widely used in signal processing applications such as video surveillance, denoising, and anomaly detection. However, these methods operate on data vectors and cannot directly be applied to higher-order data arrays. This talk introduces a new tensor-based recursive robust PCA algorithm that preserves the multidimensional structures of data. Our algorithm is based on the recently proposed tensor singular value decomposition (T-SVD). We demonstrate its operation on the application of background/foreground separation in a video stream. The background component is modeled as a gradually changing low-rank subspace. The foreground component is modeled as a sparse signal with a tensor dictionary outside the subspace. Extensive experiments on real-world videos are presented. Results indicate that our method can effectively separate background and foreground in a video.