X-Ray Video-Based 3D Bone Motion Tracking
Center for Limb Loss and Mobility
Project Goal
The goal of this work was to conduct quality research that aims to reduce limb loss and develop novel techniques using X-rays, computer vision, and image processing techniques in order to better evaluate lower limb bone motion
Challenges
This is novel technology so there is no off-the-shelf software or hardware solution available currently for these systems. Additionally, data generated from these experiments come at high computational cost due to its high dimensionality. Finally, improving the image quality and automated processing pipeline would increase the accuracy and throughput of the system.
My Solution
Automated digital signal processing and GUIs were implemented to the X-ray video tracking and analysis protocols. These custom tools reduced processing time and provided quick visualization for quality control. Additionally, mechanical design projects were conducted for unique laboratory equipment needs, including but not limited to, high-speed camera gantries and computer vision calibration objects.
Notable Features & Accomplishments
- Helped develop a software pipeline for efficiently processing high volumes of X-ray video data and recovering the 3D object poses
- Controlled the hardware using a custom LabVIEW program that synchronizes X-ray firing, camera triggering, and sensor data collection.
- Learned about image processing techniques such as flatfield correction, deconvolution, and polynomial distortion correction algorithms to correct spatial distortion, intensity roll-offs, and defocusing blurs which improved object tracking.
Skills Used
- LabVIEW
- MATLAB
- Computer Vision
- Video Tracking
- 3D Pose & Motion Estimation
- Mechanical Design
- Machining
- Prototyping
- Signal Processing
- Data Processing & Analysis
- GUI Design
- Outlier Detection
- Biomechanics
- Data Collection
- Data Acquisition
- Segmentation
- Medical Imaging
- Research
- CAD
- Calibration
- Motion Capture
Citation
Schematic image used on this page is adapted from "Marker-based validation of a biplane fluoroscopy system for quantifying foot kinematics" by Joseph Iaquinto et al.