Scanlite Analysis ROS2 Package

This package provides tools for real-time ultrasound bone segmentation, 3D reconstruction, ROS integration, and interactive GUI visualization.

📁 Project Structure

scanlite_analysis_ros2/
├── launch/
│   ├── odom_launch.py
│   ├── vicon_launch.py
│   └── segmentation.launch.xml
├── scanlite_analysis_ros2/
│   ├── segmentation_node.py
│   ├── reconstruction_node.py
│   ├── reconstruction_node_vicon.py
│   ├── transformation.py
│   ├── transformation_odom.py
│   ├── rosbagtest.py
│   ├── LiveDemoTool.py
│   ├── DataAcqTool.py
│   └── PyCATMAUS/
│       ├── SegBone.py
│       └── TransFunction.py
├── package.xml
└── CMakeLists.txt

ROS Nodes

segmentation_node.py

• Publishes real-time segmentation results as ROS topics.
• Independent from GUI, facilitating integration into broader ROS systems.

reconstruction_node.py and reconstruction_node_vicon.py

• Performs real-time 3D reconstruction of segmented bone surfaces
• Supports both IMU and Vicon motion tracking systems
• Subscribes to:
  • /us_image for ultrasound images
  • /odomimu_transform or /rigid_body_transforms for motion tracking
• Publishes:
  • Segmented images (/seg_bone/image)
  • 3D point cloud data (/rec_bone/points)
• Features interactive 3D visualization using Matplotlib
• Includes coordinate axes visualization for scanner position tracking

transformation.py

• Converts motion capture data to ROS transforms and visualization markers
• Features:
  • Coordinate system conversion for RViz visualization
  • TF broadcasting for rigid body transforms
  • Visualization markers for tracking visualization
  • Configurable parameters for different mocap systems
• Publishes:
  • /visualization_marker_rb for RViz visualization
  • /rigid_body_transforms for transform data
  • TF frames for each rigid body

transformation_odom.py

• Converts Odometry data to ROS transforms and visualization markers
• Features:
  • Real-time conversion of odometry to transform messages
  • Visualization markers for tracking visualization
  • TF broadcasting for odometry transforms
• Publishes:
  • /visualization_marker_odom for RViz visualization
  • /odomimu_transform for transform data
  • TF frames for odometry tracking

Launch Files

To run the system, you can use the following launch commands:

For Vio-based tracking:

ros2 launch scanlite_analysis_ros2 odom_launch.py

This launches:

• transformation_odom.py for IMU data processing
• reconstruction_node.py for 3D reconstruction

For Vicon-based tracking:

ros2 launch scanlite_analysis_ros2 vicon_launch.py

This launches:

• transformation.py for Vicon data processing
• reconstruction_node_vicon.py for 3D reconstruction

🖥 GUI Tools (Tkinter)

Tool	Description
DataAcqTool.py	Advanced GUI tool for real-time data acquisition and playback. Features include: - Real-time data recording - Frame-by-frame playback - Data export in multiple formats (PNG, MAT) - Segmentation parameter tuning
LiveDemoTool.py	Enhanced real-time bone segmentation visualization tool with: - Interactive 3D reconstruction - Real-time scanner tracking visualization - Data recording and export capabilities - Segmentation parameter controls

Functional Overview

ROS Nodes

• segmentation_node.py
  • Publishes segmentation results independently via ROS topics.

ROS Bag Testing (CatMausApp)

• rosbagtest.py
  • Select and play ROS bags for validating segmentation.

Features:

• ROS Integration:

  • Subscribes to /us_image and motion tracking topics
  • Supports both IMU (/odomimu_transform) and Vicon (/rigid_body_transforms) tracking
  • Synchronizes motion data with ultrasound images

• Graphical User Interface (GUI):

  • Developed with Tkinter
  • Displays ultrasound images with segmentation overlays using Matplotlib
  • Interactive controls for segmentation parameters (F0, F1, Bth, JC)
  • Real-time 3D visualization with coordinate axes

• Bone Segmentation:

  • Real-time segmentation via the BoneSeg() method
  • Parameter tuning for optimal results

• 3D Reconstruction:

  • Converts segmented 2D data into interactive 3D visuals
  • Supports multiple motion tracking systems
  • Features interactive zooming and reset functionalities
  • Real-time point cloud publishing for RViz visualization

• Data Management:

  • Real-time data recording
  • Frame-by-frame playback
  • Multiple export formats (PNG, MAT)
  • Data visualization and analysis tools

Explanation of Segmentation Parameters

Parameter	Meaning	Effect on Segmentation
F0	Energy continuity weight	Controls smoothness along bone surfaces. Higher values enforce smoother edges.
F1	Energy smoothness weight	Penalizes abrupt depth changes. Higher values produce gradual transitions.
Bth	Bone threshold	Pixel intensity threshold. Lower values include more pixels as bone.
JC	Jump constraint	Limits allowed segmentation path jumps. Higher values increase path flexibility.

Usage

Launch the ROS segmentation nodes and related topics:

ros2 launch scanlite_analysis_ros2 segmentation_launch.py 

Run the 3D reconstruction node (choose based on tracking system):

# For Vio tracking
ros2 run scanlite_analysis_ros2 reconstruction_node.py

# For Vicon tracking
ros2 run scanlite_analysis_ros2 reconstruction_node_vicon.py

Start live device acquisition GUI:

ros2 run scanlite_analysis_ros2 segmentation_node.py

Real-time segmentation visualization tool:

ros2 run scanlite_analysis_ros2 LiveDemoTool.py

Requirements

• ROS2 Humble
• Python 3.x
• Tkinter
• Matplotlib
• NumPy
• OpenCV
• SciPy

---

This project relies on PyCATMAUS, a library used for segmentation and transformation functions.

🔗 Repository: PyCATMAUS GitHub