Human Body Detection with Occlusion
Session Number
Project ID: CMPS 11
Advisor(s)
Dr. Changliu Liu
Mr. Rui Chen, Carnegie Mellon University, Robotics Institute, Intelligent Control Lab
Discipline
Computer Science
Start Date
17-4-2024 8:55 AM
End Date
17-4-2024 9:10 AM
Abstract
The purpose of this design experiment was to attempt to determine the accuracy/possibility of using extended Kalman filters (EKFs) to approximate a human’s shoulder’s location when occluded in a depth camera’s point of view. The project was conducted entirely through code and avoided any human involvement/error. The outcome of this research project may result in more accurate measurements for human-body detection, potentially allowing robots to contribute to human goals without interference in real human-robot interaction environments.
This project consists of the following two different procedures: 1) design steps of solution, 2) testing final design iteration. To construct the solution to the project, the first procedure must be followed, but to test the solution, the second procedure must be followed. Each procedure is extremely descriptive, detailing the several steps to building, programming, and analyzing the design solution.
In conclusion, the problem statement was ultimately supported. Throughout all the data collected and graphs constructed, the extended Kalman filter’s average mean-squared errors were consistently and significantly lower than those of the simulated observation (camera- perceived) data. The project’s findings reveal the potential for the extended Kalman filter in real-world environments and its application to keypoints across the human body.
Human Body Detection with Occlusion
The purpose of this design experiment was to attempt to determine the accuracy/possibility of using extended Kalman filters (EKFs) to approximate a human’s shoulder’s location when occluded in a depth camera’s point of view. The project was conducted entirely through code and avoided any human involvement/error. The outcome of this research project may result in more accurate measurements for human-body detection, potentially allowing robots to contribute to human goals without interference in real human-robot interaction environments.
This project consists of the following two different procedures: 1) design steps of solution, 2) testing final design iteration. To construct the solution to the project, the first procedure must be followed, but to test the solution, the second procedure must be followed. Each procedure is extremely descriptive, detailing the several steps to building, programming, and analyzing the design solution.
In conclusion, the problem statement was ultimately supported. Throughout all the data collected and graphs constructed, the extended Kalman filter’s average mean-squared errors were consistently and significantly lower than those of the simulated observation (camera- perceived) data. The project’s findings reveal the potential for the extended Kalman filter in real-world environments and its application to keypoints across the human body.