SSIM Renderer Library

SSIM Renderer is a Qt-based C++ library for OpenGL enabled rendering of virtual X-Ray images from statistical shape and intensity models.

Description

The main focus of the library is a rendering of virtual X-Ray images from a statistical shape and intensity model (SSIM) proposed by R. Yao [1]. The library provides OpenGL accelerated rendering of the SSIM according to [2]. Moreover, it includes OpenGL and OpenCL implementations of two image similarity metrics: joint histogram normalized mutual information metric and sum of squared differences metric. The library is highly optimised and suitable for purposes of 2D-3D deformable registration [3]. It is also capable to work with simple polygonal statistical shape models (SSM) stored in Statismo file format.

Version from 17 December 2015. Updated on 26 January 2023.

Features

• Rendering of virtual X-Ray images from SSIM.
• Rendering of density images with float values from SSIM.
• Rendering of surface images from SSIM/SSM.
• Rendering of silhouettes of SSIM/SSM.
• Mirroring of the shape models.
• Sharing shape model between many renderers using OpenGL shared contexts.
• Exporting the surface of the shape model in STL file format.
• Computation of OpenGL and OpenCL accelerated image similarity metrics.
• etc.

Installation

The project requires following toolkits and libraries installed:

• Qt 5 or later
• HDF5
• libmeshb7 (https://github.com/LoicMarechal/libMeshb)
• matio (https://github.com/tbeu/matio)
• zlib
• OpenEXR (https://github.com/AcademySoftwareFoundation/openexr)

Envriroment variables must be set according to ssimrenderer_dependents.pri file.

The library has been developped for usage with Windows 7 or later.

Examples

To get started with the SSIM Renderer library, please have a look at the included examples:

• ImageMetrics.cpp

  • example demonstrating computation of image similarity metrics between two virtual X-Ray images

• SimpleStatismoModel.cpp

  • demonstration of the offscreen rendering of a simple polygonal shape model

• IntensityShapeModel.cpp

  • example of rendering virtual X-Ray images from a statistical shape and intensity model
    
  Digitally reconstructed radiograph rendered from a statistical shape and intensity model.

• DensityImage.cpp
  • example shows how to render a density image containing float values. The resulting image is stored in OpenEXR format (simple viewer is available at https://github.com/afichet/openexr-viewer). The example allows to render images in physical units.
    
    
  Screen shot of a density image in the OpenEXR viewer.

There is also a full reference manual available.

Downloading

The SSIM Renderer library can be further redistributed under the terms of the LGPL version 3 open source license. The library can be obtained from the following location: http://www.fit.vutbr.cz/research/prod/?id=458

Citation

If you use the library, please cite the following research:

Ondrej Klima, Petr Kleparnik, Michal Spanel, and Pavel Zemcik "Intensity-based femoral atlas 2D/3D registration using Levenberg-Marquardt optimisation", Proc. SPIE 9788, Medical Imaging 2016: Biomedical Applications in Molecular, Structural, and Functional Imaging, 97880F (29 March 2016); https://doi.org/10.1117/12.2216529

Article fulltext in SPIE Digital Library

Acknowledgment

This work has been supported by the Technology Agency of the Czech Republic (TA CR, Project Id: TA04011606).

Authors

• Petr Kleparnik ([email protected])
• Ondrej Klima ([email protected], ORCID: 0000-0001-9295-065X, https://www.fit.vut.cz/person/iklima/.en, https://www.researchgate.net/profile/Ondrej-Klima-4)
• Michal Spanel ([email protected])
• Pavel Zemcik ([email protected])

References

[1] J. Yao, R. Taylor, "Construction and simplification of bone density models" SPIE Medical Imaging: 2001.

[2] M. Ehlke, HRamm, H. Lamecker, H.C. Hege, S. Zachow. "Fast generation of virtual X-ray images for reconstruction of 3D anatomy." IEEE Trans Vis Comput Graph: Dec 2013.

[3] O. Klima, P. Kleparnik, M. Spanel, P. Zemcik, "GP-GPU accelerated intensity-based 2D/3D registration pipeline" In Proceedings of Shape Symposium, p. 19, Delemont, Switzerland, 2014.