PTV-Flow-Analysis-Endovascular-Simulator

This repository contains a MATLAB implementation of Particle Tracking Velocimetry (PTV) for analyzing fluid flow in an endovascular simulator. The system tracks tracer particles in high-speed video sequences to estimate velocity fields, enabling quantitative hemodynamic analysis in vascular models.

Results Summary

Validation

The system was validated against known reference flow rates:

Reference Flow (mL/min)	Estimated Flow (mL/min)	Error
775	853	+10.1%
1,700	1,764	+3.8%

Key Features

• Automated Particle Detection: HSV color filtering with morphological cleaning
• Robust Tracking: Hungarian algorithm-based particle matching with displacement constraints
• Flow Quantification: Velocity field estimation and volumetric flow rate calculation
• Interactive Calibration: User-guided pixel-to-mm conversion and ROI selection
• Comprehensive Visualization: Vector fields and speed histograms
• Video Output: Animated velocity fields and particle masks

Getting Started

Prerequisites

• MATLAB R2020a or later
• Required toolboxes:
  • Image Processing Toolbox
  • Statistics and Machine Learning Toolbox

Running the Pipeline

Step 1: Place your video file in the project directory

Step 2: Update the configuration parameters in src/ptv_flow.m

Step 3: Run the main script

ptv_flow();

Configuration

Modify the following parameters in src/ptv_flow.m

Parameter	Description	Default Value
videoFile	Input video filename	'10V_slow.mp4'
tubeWidth	Tube diameter for flow calculation	10 mm
dt	Frame rate of slow-motion recording	1 / 120 s
minParticleArea	Minimum particle area (pixels)	8
minCircularity	Particle circularity threshold	0.7
maxDisplacement	Max displacement between frames	50 pixels

Repository Structure

├── data/
│ ├── 10V_slow.mp4            # Slow Flow Video
| └── 16V_slow.mp4            # High Flow Video
├── results/                  # Results generated automatically
│ ├── 10V_Particle_Masks.mp4  # Animated video of detected particle masks
│ ├── 10V_Velocity_Field.mp4  # Animated velocity field visualization
│ ├── 10V_Vector_Field.png    # Complete velocity vector field visualization
│ ├── 10V_Speed_Histogram.png # Speed distribution histogram with velocity components for Slow Flow
│ └── 16V_Speed_Histogram.png # Speed distribution histogram with velocity components for High Flow
├── src/
│ └── ptv_flow.m              # Main PTV analysis script
└── README.md                 # This file

Output Files

After running the pipeline, you'll find:

File	Description
Vector_Field.png	Complete velocity vector field visualization
Speed_Histogram.png	Speed distribution histogram with velocity components
Particle_Masks.mp4	Animated video of detected particle masks
Velocity_Field.mp4	Animated velocity field visualization

Methodology

1. Video Preprocessing

• Convert frames to HSV color space
• Apply color thresholding for particle isolation
• Morphological operations for noise removal
• Circularity-based particle filtering

2. Particle Tracking

• Centroid detection using regionprops
• Cost matrix construction (Euclidean distances)
• Hungarian assignment algorithm for optimal matching
• Displacement thresholding for outlier rejection

3. Flow Calculation

• Pixel-to-mm conversion via user calibration
• ROI-based velocity extraction
• Volumetric flow rate calculation using cross-sectional area
• Statistical analysis of velocity components

License

This project is licensed under the MIT License - see the LICENSE file for details.