[CVPR2023] Learning Partial Correlation based Deep Visual Representation for Image Classification
Rahman, Saimunur, Piotr Koniusz, Lei Wang, Luping Zhou, Peyman Moghadam, and Changming Sun. "Learning partial correlation based deep visual representation for image classification." In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 6231-6240. 2023.
Visual representation based on covariance matrix has demonstrates its efficacy for image classification by characterising the pairwise correlation of different channels in convolutional feature maps. This paper proposes an Iterative method to solve Sparse Inverse Covariance Estimation (iSICE). Our work obtains a partial correlation based deep visual representation and mitigates the small sample problem often encountered by covariance matrix estimation in CNN.
Website: https://csiro-robotics.github.io/iSICE/
Open Source Code: https://github.com/csiro-robotics/iSICE
Paper: http://openaccess.thecvf.com/content/...
Research collaboration between CSIRO's Embodied AI Cluster, University of Wollongong , Australian National University, University of Sydney, QUT (Queensland University of Technology)