MRI and CT 3D Viewer

# DICOM MRI & CT 3D Reconstruction and Medical Imaging Viewer ## 1. Project Overview Create a professional Windows desktop application named **Medical 3D Imaging Viewer**. The application will import, process, visualize, and reconstruct medical imaging data from MRI (Magnetic Resonance Imaging) and CT (Computed Tomography) scans stored in DICOM (Digital Imaging and Communications in Medicine) format. The main objective is to transform a series of 2D medical image slices into an interactive, anatomically accurate 3D representation of the scanned body region. The application should support imaging of the pelvis, abdomen, chest, brain, spine, bones, organs, and other anatomical structures. Users should be able to examine individual DICOM slices, generate volumetric 3D visualizations, isolate anatomical structures, measure distances, and export reconstructed 3D models. The application is intended for medical image visualization, educational purposes, research, and technical analysis. It is not intended to independently diagnose disease or replace a radiologist. ## 2. Development Technologies Build the application using the following technologies: - **Programming language:** C++17 or C++20 - **IDE:** Microsoft Visual Studio 2022 - **Operating system:** Windows 10/11, 64-bit - **GUI framework:** Qt 6 - **Medical image processing:** ITK (Insight Toolkit) - **3D visualization:** VTK (Visualization Toolkit) - **DICOM support:** GDCM and ITK DICOM readers - **Build system:** CMake - **Dependency management:** vcpkg - **Optional GPU acceleration:** VTK OpenGL GPU volume rendering - **Export formats:** STL, OBJ, PLY, VTI, NIfTI, and supported medical imaging formats Use a modular architecture to separate image loading, processing, segmentation, reconstruction, visualization, measurements, and file export. ## 3. DICOM File Import The application must support importing complete DICOM studies containing hundreds or thousands of medical image slices. Supported input options: 1. Select a folder containing DICOM files. 2. Select multiple DICOM files manually. 3. Import a DICOMDIR study. 4. Import supported compressed DICOM datasets. 5. Import CT and MRI studies independently. The application should automatically: - Identify DICOM files. - Read patient and study metadata. - Group images by Study Instance UID and Series Instance UID. - Identify the imaging modality (MRI or CT). - Determine image orientation, pixel spacing, and slice geometry. - Sort image slices using DICOM spatial metadata rather than filenames. - Detect missing, duplicated, or inconsistent slices. - Assemble compatible image slices into a correctly oriented 3D volume. Display available studies and series in a navigation panel. Do not combine different MRI sequences or incompatible imaging series into a single volume automatically. ## 4. 2D Medical Image Viewer Provide a medical image viewing interface with three synchronized anatomical planes: - **Axial:** Horizontal cross-sectional view. - **Sagittal:** Side view. - **Coronal:** Front view. Users should be able to: - Scroll through individual image slices. - Zoom in and out. - Pan images. - Adjust brightness and contrast. - Adjust window width and window level. - Rotate images. - Display crosshair coordinates. - View image orientation labels. - Show slice position and thickness. - Measure distances in millimeters. - Compare different imaging series. When the user selects a point in one anatomical plane, the corresponding position should be synchronized in the other planes. ## 5. Automatic 3D Volume Reconstruction The primary feature of the application is converting 2D DICOM image slices into an interactive 3D volume. The reconstruction pipeline should: 1. Read the selected DICOM image series. 2. Validate spatial metadata and image consistency. 3. Reconstruct the volumetric image dataset. 4. Apply appropriate interpolation when needed. 5. Preserve original physical dimensions and voxel spacing. 6. Generate a 3D visualization using VTK. 7. Display the result in an interactive 3D viewport. The application must support two reconstruction methods. ### A. Volume Rendering Display the complete image volume using GPU-accelerated rendering. Allow the user to adjust: - Opacity - Color mapping - Transfer functions - Window and level - Clipping planes - Rendering quality - Camera angle ### B. Surface Reconstruction Create polygonal 3D surfaces from segmented anatomical structures. Use suitable methods such as Marching Cubes or Flying Edges. The user should be able to rotate, zoom, pan, isolate, hide, or display reconstructed structures. The 3D model should preserve the physical geometry of the original medical scan. ## 6. CT Image Reconstruction For CT studies, support Hounsfield Unit (HU) calibration using the DICOM rescale parameters. Provide adjustable visualization presets for: - Bone - Soft tissue - Lung - Blood vessels when contrast imaging is available - Other suitable CT tissue ranges Support threshold-based segmentation where appropriate. For example, the user should be able to isolate the pelvic bones from a CT scan and create a 3D skeletal model. Do not assume that thresholding alone can reliably distinguish every organ, tumor, or soft-tissue structure. ## 7. MRI Image Reconstruction For MRI studies, support different MRI sequences, including: - T1-weighted - T2-weighted - Diffusion-weighted imaging (DWI) - ADC maps - Contrast-enhanced MRI - Other supported sequences The application should identify and display available MRI sequences separately. Provide image enhancement, noise reduction, intensity normalization, and contrast adjustment options. Because MRI intensities are not standardized like CT Hounsfield Units, segmentation should use sequence-appropriate methods rather than universal fixed thresholds. Support 3D reconstruction of selected anatomical regions when image resolution and segmentation quality are sufficient. ## 8. Anatomical Segmentation Provide tools for isolating anatomical structures from MRI or CT images. Segmentation tools should include: - Manual drawing - Brush and eraser - Region growing - Threshold segmentation - Connected-component analysis - Morphological processing - Slice-by-slice editing - Optional AI-assisted segmentation Allow users to create multiple labeled structures with different colors. For example: - Bones: white - Organs: different selectable colors - Blood vessels: red - User-defined region of interest: yellow The application should display segmented structures in both 2D slices and the 3D viewer. Users must be able to review and correct segmentation results before exporting or measuring structures. AI-assisted segmentation should be optional and clearly indicate uncertainty and limitations. ## 9. Interactive 3D Viewer Create a professional 3D viewer with the following capabilities: - Rotate the reconstructed anatomy 360 degrees. - Zoom and pan. - Reset the camera. - Switch between perspective and orthographic views. - Show or hide individual structures. - Change colors and opacity. - Display an anatomical orientation marker. - Cut through the volume using clipping planes. - Display 2D slices intersecting the 3D volume. - Save screenshots. - Export 3D models. The user should be able to view internal structures by making outer structures transparent or hiding them. ## 10. Measurements and Analysis Implement measurement tools using the physical spatial information stored in DICOM metadata. Supported measurements should include: - Point-to-point distance in millimeters. - Angle measurement. - Area measurement in square millimeters. - Segmented structure volume in cubic centimeters. - Three-dimensional bounding-box dimensions. - Coordinates of selected points. - Distance between selected anatomical landmarks. Measurements should account for voxel spacing, image orientation, and reconstructed geometry. Display a warning if the required spatial calibration information is missing or unreliable. ## 11. Multiple Scan Comparison Allow users to load more than one medical imaging study. Examples include: - CT and MRI of the same anatomical region. - Two MRI studies from different dates. - Pre-treatment and post-treatment scans. Provide side-by-side comparison and synchronized navigation. Include optional image registration to align datasets acquired at different times or using different modalities. Registration should use suitable ITK algorithms and allow the user to inspect alignment quality. Do not assume that CT and MRI images will align automatically. ## 12. Main Application Interface Design a modern, professional interface with a dark medical imaging theme. The main application window should include: **Left Panel – Study Explorer** Display imported patients, studies, series, and available images. **Center Panel – Medical Imaging Workspace** Display four synchronized viewports: - Top left: Axial view - Top right: Sagittal view - Bottom left: Coronal view - Bottom right: Interactive 3D reconstruction **Right Panel – Tools and Settings** Include controls for segmentation, opacity, color, rendering, measurements, and visibility. **Bottom Panel – Status and Progress** Display loading progress, reconstruction progress, voxel dimensions, imaging modality, and processing messages. The layout should support resizing, maximizing viewports, and full-screen 3D viewing. ## 13. Export and Save Features Users should be able to save their work and export results. Support: - 3D models as STL, OBJ, or PLY. - Medical image volumes as NIfTI or supported VTK formats. - Segmentation masks in suitable medical imaging formats. - PNG screenshots. - Measurement reports as CSV. - Project files containing study references, segmentation data, measurements, and visualization settings. The application must preserve the original DICOM files without modifying them. Provide an option to anonymize patient information before exporting or sharing data. ## 14. Performance and Stability The application should efficiently process large DICOM studies. Requirements: - Use background threads for file loading and image processing. - Keep the interface responsive during reconstruction. - Use GPU acceleration where supported. - Provide memory usage monitoring. - Support large volumetric datasets. - Allow processing cancellation. - Handle corrupted or incomplete DICOM series gracefully. - Report missing dependencies and unsupported DICOM transfer syntaxes clearly. Use appropriate memory management and avoid unnecessary copies of large image volumes. ## 15. Application Development Phases ### Phase 1 – Core DICOM Viewer Build a working Windows application capable of importing DICOM studies, identifying MRI and CT series, and displaying axial, sagittal, and coronal slices. ### Phase 2 – 3D Reconstruction Implement volumetric reconstruction using ITK and VTK, with GPU rendering, rotation, zoom, opacity, and clipping tools. ### Phase 3 – Segmentation Add manual and semi-automatic segmentation and 3D surface generation. ### Phase 4 – Measurements and Export Implement calibrated measurements, volume calculations, screenshots, and 3D model export. ### Phase 5 – Advanced Features Add multi-study comparison, image registration, optional AI-assisted segmentation, and performance optimization. Each phase must produce a working, testable application before moving to the next phase. ## 16. Cursor AI Implementation Instructions Start by examining the existing project and installed development dependencies. If an existing C++/Qt/ITK/VTK application is present, extend it instead of rebuilding it unnecessarily. Follow these development requirements: 1. Use Visual Studio 2022 and CMake. 2. Organize the source code into clear modules. 3. Reuse compatible installed libraries. 4. Avoid downloading or rebuilding unnecessary dependencies. 5. Implement one working feature at a time. 6. Build and test the application after major changes. 7. Use sample or synthetic DICOM datasets for validation. 8. Verify that reconstructed dimensions match known physical dimensions. 9. Provide clear error messages. 10. Document build instructions and required dependencies. Do not create a simulated 3D image from a single 2D screenshot and present it as a medically accurate reconstruction. The 3D reconstruction must be based on the actual DICOM voxel data and spatial metadata. ## 17. Final Project Goal The completed application should allow a user to: 1. Open the application. 2. Import a complete MRI or CT DICOM study. 3. View all available image slices. 4. Select an imaging series. 5. Click **Generate 3D**. 6. View and rotate the reconstructed anatomical volume. 7. Isolate selected anatomical structures. 8. Perform measurements. 9. Save or export the results. **Final objective:** Develop a powerful, expandable Windows medical imaging application that transforms MRI and CT DICOM datasets into accurate, interactive 3D visualizations for technical analysis, research, and educational use.