Multiple Moving Targets Tracking based on Kernel Localization and Group Trackers for Envisioned Functional Microwave Brain Imaging Applications

This paper presents a new concept of multiple targets tracking using hierarchical trackers based on kernel localization for envisioned functional microwave brain imaging application. For this purpose, the performance of brain activated regions tracking using fMRI video is improved. In the first stage, all of the moving regions in the fMRI video frame are detected. Then, by using the group tracking, histogram and distance corresponding to the moving-targets in the previous frame, the directions of moving-regions are determined. After determining the exact number of moving areas and tracking them, one by one, the direction of each vector is extracted. In addition, due to the kernel labelling, the proposed method has capability of separation and merging by group tracking in conditions of motion paths interfere with each other. The simulated results validate the effectiveness of the proposed method for precisely tracking of the activated regions.