High Precision Partial Object Tracking using Intensity and Depth Data

Authors

Presenter(s)

Eric G. Smith

Comments

Presentation: 11:40-12:00, Kennedy Union 222

Files

Description

Object and target tracking algorithms often have scenes and objects that they are better utilized for. However, the goal for object tracking algorithms is to be robust enough to be employable in many scenarios with as few disadvantages as possible. This project attempts to leverage open-source object tracking algorithms and combine the tracking performance of each for improved tracking capabilities. This fusion approach is done utilizing OpenCV, an open-source library for real-time computer vision functionality. An image set with objects of interest is used as the data source. The performance of individual trackers will be analyzed and compared to the performance of the fusion approach this project attempts to leverage. The goal of this project is to leverage the capabilities of each tracker and fuse their track results in a way to make up for poor performance in each algorithm individually. The resulting algorithm tracks a part of the object with sub-pixel precision.

Publication Date

4-17-2024

Project Designation

Graduate Research

Primary Advisor

Yakov Diskin, K. Asari Vijayan

Primary Advisor's Department

Electrical and Computer Engineer

Keywords

Stander Symposium, School of Engineering

Recommended Citation

"High Precision Partial Object Tracking using Intensity and Depth Data" (2024). Stander Symposium Projects. 3422.
https://ecommons.udayton.edu/stander_posters/3422