AEHRC has developed a flexible surgical simulation software platform, MILXSim, which provides high performance rendering and simulation capabilities to the Surgical Simulation and Planning team. The platform is developed in C++ and uses OpenGL with advanced surface shading techniques to render photo-realistic surgical training scenes and data rich preoperative planning scenarios. MILXSim uses a plugin architecture to simplify development and deployment of new features and reduces the complexity of developing multi-threaded simulation scenes with a rich configuration system. The platform exposes the power of the modern Graphics Processing Unit (GPU) using the NVIDIA CUDA programming API along with built-in support for surface and volumetric meshing.
By developing a plugin-driven architecture for surgical simulation, core features of the simulation platform are immediately available for use in new projects and novel purposes. The decision to use industry standard OpenGL rendering platform and NVIDIA CUDA for high performance GPGPU programming ensures that MILXSim provides a high level of performance with photo-realistic rendering capabilities at interactive rates. The multi-threaded architecture and support for inter-thread communication frees developers from the complexity inherent in multi-threaded systems and allows efforts to be centred on novel research. Support for common haptic devices is built in and adding new devices via the plugin architecture requires minimal effort.
Support for real-time surface shading using GLSL is provided via a rich material system. Integrated support for advanced 2D overlays, multiple cameras, lights and view ports ensures that MILXSim is capable of handling challenging GUI designs. A fully featured XML configuration system allows interactive GUI controls to be developed with values wired directly to OpenGL GLSL shaders.
MILXSim is currently used for all Surgical Simulation and Preoperative Planning projects. Currently it forms the basis for CSIRO's Colonoscopy Simulator as well as our interactive 3D image analysis tool, Vulture, and was used to prototype our novel techniques for real-time non-linear tissue deformation and free hand ultrasound simulation.
Last Updated on Friday, 23 September 2011 09:30