Medical images are visual representations of interior structures and functions of a human body. These images play a vital role in diagnosis, monitoring and treating medical conditions.
Medical imaging modalities such as Computed Tomography (CT) and Magnetic Resonance Imaging (MRI) generate large volumetric datasets composed of stacked 2D slices that form a 3D voxel-based volume. Analyzing these datasets through traditional slice-by-slice inspection can be difficult and time-consuming.
This project introduces a medical visualization prototype that integrates computer graphics and medical imaging techniques to assist in tumor analysis.
The system provides the following capabilities:
- Conversion of segmented medical imaging data into interactive 3D tumor models
- 2D slice navigation for detailed inspection
- Interactive exploration of tumor structures
- Quantitative tumor volume calculation
The objective of this prototype is to demonstrate how scientific visualization techniques can assist healthcare professionals in understanding tumor structure, spatial relationships, and progression over time.
Note:
For a detailed Software Engineering specification, please refer this Technical Report
This section describes the technologies, design decisions, and implementation methods used to build the system.
The prototype was developed using several technologies and libraries, listed bellow.
VTK is an open-source software system widely used for scientific visualization, computer graphics, and image processing. It was originally developed by Will Schroeder, Ken Martin, and Bill Lorensen to provide a unified platform for visualization research and development.Including, it provides powerful tools for 2D and 3D visualization
Source:
Qt is a cross-platform framework for graphical user interface (GUI) development. It provides a rich collection of widgets, layout systems, and UI components that allow developers to build complex and interactive applications.
Source:
There are two main approaches for visualizing volumetric medical data:
- Volume Rendering
- Surface Rendering
In this project, surface rendering was chosen because it provides a clear and intuitive visualization of tumor boundaries and spatial relationships with surrounding anatomical structures.
One of the most widely used algorithms for surface rendering is the Marching Cubes algorithm, which extracts an isosurface from volumetric data.
Reference:
- Lorensen, W. E., & Cline, H. E. (1987).
Marching Cubes: A High Resolution 3D Surface Construction Algorithm.
https://doi.org/10.1145/37401.37422
The system supports several features designed to improve medical data visualization and tumor analysis.
Users can upload medical imaging datasets in:
- DICOM
- NIfTI
Once loaded, the system automatically displays the dataset and enables navigation through the slices.
The system provides interactive slice navigation and supports:
- Axial
- Coronal
- Sagittal
views through Multi-Planar Reconstruction (MPR), enabling users to inspect anatomical structures from multiple perspectives.
The system allows users to upload segmentation masks corresponding to the medical dataset.
Features include:
- Overlaying segmentation masks on 2D slices
- Rendering segmented tumor regions in 3D
- Assigning distinct colors to different labels
- Selecting which tumor regions to display
The system computes the volume for each tumor region using the segmentation data.
The system supports dynamic contrast enhancement (windowing).
Users can manually adjust:
- Brightness
- Contrast
This improves the visibility of anatomical structures and helps emphasize specific tissues.
Users can dynamically adjust:
- Opacity — controls transparency of the 3D model
- Iso-value — determines the threshold used for surface extraction
These controls allow users to visualize different anatomical structures more effectively.
One of the most important features of the system is the ability to compare two tumors.
The system provides:
- 2D comparison
- 3D comparison
- Volume difference calculation
This feature is useful for:
- Tracking tumor progression
- Evaluating treatment response
- Comparing tumors from different scans or patients
The interface supports both:
- Light mode
- Dark mode
to improve visual comfort and usability.
The system was tested using datasets obtained from both local medical institutions and international medical repositories.
The datasets included different imaging modalities such as:
- MRI
- CT
- X-ray
and tested both:
- DICOM format
- NIfTI format
One of the global datasets used for testing was obtained from:
The Cancer Imaging Archive (TCIA)
[https://www.cancerimagingarchive.net]
User testing (acceptance testing) was conducted to ensure the system is usable, intuitive, and suitable for medical professionals.
Feedback sessions were held with:
- Medical academy members
- Oncologists
- Radiologists
Their feedback helped improve the system’s usability, interface design, and functionality.
This project successfully demonstrates a medical visualization tool that integrates:
- 2D medical image visualization
- Interactive 3D tumor rendering
- Segmentation visualization
- Tumor volume calculation
- Tumor comparison tools
Future improvements may include:
- Built-in tumor segmentation
- Tumor classification using machine learning
- Interactive measurement tools
- Additional clinical analysis features
The long-term goal is to develop advanced visualization tools that support clinicians in understanding complex medical data and making more informed decisions.
Main author:
Sarah Abu Irmeileh
- Computer Science – Palestine Polytechnic University
Supervisor:
Dr. Zein Salah
- Assistance Professor at the College of Computer Engineering and Information Technology – Palestine Polytechnic University
This project is intended for academic and research purposes, refer to it with proper citation.