Robot tracker based on semantic segmentation computer vision algorithms

Jovanov, Aleksandar (2017) Robot tracker based on semantic segmentation computer vision algorithms. In: PROCEEDINGS of the 14th Conference on Informatics and Information Technology. Faculty of Computer Science and Engineering, Ss. Cyril and Methodius University in Skopje, Macedonia, Skopje, Macedonia, pp. 218-221. ISBN 978-608-4699-07-1


Download (146kB) | Preview
Official URL:


This project presents an approach to human tracking based on a semantic segmentation algorithm which uses convolutional neural networks and optical flow based on Farneback’s method. Tracking is implemented via person discrimination using Neural Networks and Speeded Up Robust Features as an initial step and optical flow tracking in next steps if movement is present. Actual robot actuation, localization and central planning are done by the move_base package of the Robot Operating System navigation stack and ROSARIA package with an intermediary node that transforms image data to goal coordinates. Microsoft Kinect and p3dx robot are the hardware components used in the realization of the project.

Item Type: Book Section
Subjects: International Conference on Informatics and Information Technologies > Students Session
Depositing User: Vangel Ajanovski
Date Deposited: 29 Nov 2017 18:27
Last Modified: 29 Nov 2017 18:27

Actions (login required)

View Item View Item