Optimization Spectral Tracker Algorithm

Fakultas Ilmu Komputer Universitas Indonesia > E-News > Optimization Spectral Tracker Algorithm


There are several factors that affect the performance of tracking algorithm. Those several factors include robust and speed. Thus, Alexander Agung Santoso Gunawan in his dissertation titled “Pelacakan Visual Cepat dan Optimal Berdasarkan Metode Spektral dalam Kerangka Bayesian” proposed a new tracking algorithm in relation to probabilistic Bayesian Framework. This algorithm proposed will be developed in order to efficiently solve optimal particle filter (OPF) by using a spectral method or spectral tracker (ST). Moreover, in his dissertation, Alex proposed to use multiple observation points at once and the information of movement observation points so that can estimate the scale and also the rotation. Through his dissertation, Alex investigated the problem of short-term model-free tracking which its main purpose tracks any objects based on one annotation box from one object that is in the first frame only. The result shows that there is a well improved performance especially in terms of robustness and speediness.

These results were presented by Alexander Agung Santoso Gunawan in an open session which was held on June 25th, 2016 in Building B’ Auditorium of Fasilkom. The Computer Science doctorate conferment session was led by Mirna Adriani, Ph.D., with Prof. T. Basaruddin, Ph.D., Dr. Petrus Mursanto, Dr. M. Ivan Fanany, Bob Hardian, Ph.D., and Dr. Agus Buono as the examiners. In the session, Prof. Dr. Aniati Murnia A. and Dr. Wisnu Jatmiko were also present as the promoter and co-promoter respectively. The dissertation achieved a Cumlaude result, and a doctorate in Computer Science was then conferred to Dr. Alexander Agung Santoso Gunawan as the 57th graduate from Fasilkom UI Doctorate program. (HK)