,
College of Computer Science and Information Technology, University of Kirkuk , Kirkuk , Iraq
,
College of Science, University of Kirkuk , Kirkuk , Iraq
,
College of Law and Political Science, University of Kirkuk , Kirkuk , Iraq
Lecturer, College of Computer Science and Information Technology, University of Kirkuk , Kirkuk , Iraq
Computer vision is a significant field of application of mathematical models developed to monitor 3D anatomical locations of the human body, particularly in robotics, surveillance, and medicine. In this paper, we present a new model that applies the Kalman filter (KF) to track 3D anatomical features in real time with increased precision. The approach separates video frames, identifies objects using the pixel characteristics, and uses the Kalman filter to forecast and correct landmark locations. As shown, experimental results demonstrate that the proposed solution is much more efficient than conventional tracking mechanisms. The system's tracking accuracy has improved, with a mean squared error (MSE) of 0.035 compared to 0.048 for the baseline Kalman filter. The suggested approach provides a processing time of 15 ms per frame, which is sufficient for real-time operation. The fact that the system can cope with noisy data and any high-speed movements makes it robust and is applicable in cases where the application needs high precision and low latency. The test confirms that Kalman filtering and deep learning feature-extraction frameworks improve the tracking accuracy and stability of 3D anatomical landmarks in dynamic systems. This study opens the way to more trustworthy real-time medical surveillance, rehabilitation, and robotic interaction tracking solutions.
This is an open access article distributed under the Creative Commons Attribution Non-Commercial License (CC BY-NC) License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
0
The statements, opinions and data contained in the journal are solely those of the individual authors and contributors and not of the publisher and the editor(s). We stay neutral with regard to jurisdictional claims in published maps and institutional affiliations.