Robust Visual Object Tracking in Clustered Environment
L. Krishna Kumari, K Ramalakshmi, V. SrinivasaRaghavan · 2022
To perform visual object tracking in clustered environment, kernel methods based Support Vector Machines (SVMs) can be used. A Scale adaptive Kernel support correlation filter Algorithm (SKSCF) is employed here for visual object tracking. In this method, the SVM models with circulant matrix formulation have been used. For visual tracking and its related applications it enhances the process of optimization. SVM models with circulant matrix formulation use the discrete Fourier transform operation to accomplish visual object tracking. The issue of visual objects is illustrated as a recursive command of Support Correlation Filters (SCFs). A Scale Adaptive Kernel Support Correlation Filter (SKSCF) has an O(n2 log n) computational complexity. The SVM-based techniques’ computational complexity is at O (n4). Along with the Scale Adaptive Kernel Support Correlation Filter Algorithm, factors like multi-channel features, scale adaptive approach, and kernel methods are applied. For a sizable standard dataset, the Scale Adaptive Kernel Support Correlation Filter Algorithm provides finer outcomes in terms of reliability and speed.