Patient-specific 3D Ultrasound Simulation Based on Convolutional Ray-tracing and Appearance Optimization
The simulation of medical ultrasound from patient-specific data may improve the planning and execution of interventions e.g. in the field of neurosurgery. However, both the long computation times and the limited realism due to lack of acoustic information
- PDF / 3,383,819 Bytes
- 9 Pages / 439.363 x 666.131 pts Page_size
- 91 Downloads / 191 Views
4
Abstract. The simulation of medical ultrasound from patient-specific data may improve the planning and execution of interventions e.g. in the field of neurosurgery. However, both the long computation times and the limited realism due to lack of acoustic information from tomographic scans prevent a wide adoption of such a simulation. In this work, we address these problems by proposing a novel efficient ultrasound simulation method based on convolutional ray-tracing which directly takes volumetric image data as input. We show how the required acoustic simulation parameters can be derived from a segmented MRI scan of the patient. We also propose an automatic optimization of ultrasonic simulation parameters and tissue-specific acoustic properties from matching ultrasound and MRI scan data. Both qualitative and quantitative evaluation on a database of 14 neurosurgical patients demonstrate the potential of our approach for clinical use.
1
Introduction
A realistic simulation of medical ultrasound is an important tool, e.g. for transducer design, training of physicians or multi-modal image-registration through simulation. A further attractive application is pre-operative planning, in which a patient-specific ultrasound simulation of the operational situs could help the surgeon to anticipate tissue appearance or optimal transducer positioning. However, a wide adoption in this context has been prevented by two problems. First, computation times of realistic ultrasound simulation methods still prevent interactive frame rates. Second, deriving the required acoustic parameters of tissue from a CT or MRI scan of the same patient is difficult due to different physical imaging principles and limited resolution of the source modalities. In this work, we are addressing both problems by proposing an interactive and realistic simulation based on convolution and ray-tracing with simulation parameters that can be optimized to match the appearance of real ultrasound images. c Springer International Publishing Switzerland 2015 N. Navab et al. (Eds.): MICCAI 2015, Part II, LNCS 9350, pp. 510–518, 2015. DOI: 10.1007/978-3-319-24571-3_61
Ultrasound Simulation
511
Ultrasound simulation approaches can be roughly categorized into wave-based, ray-based and convolution-based methods. Wave-based methods offer the highest realism and physical accuracy due to actual simulation of wave-front propagation in tissue. However, they are computationally expensive, requiring up to one hour for rendering of a single frame even on modern graphic card hardware [9]. Another approach is to simulate the spatial impulse response of the ultrasound system and convolve it with an artificial map of micro-scatters. A well-known software to employ this model is Field II, which takes up to one minute for simulation of a 2D image [8] and often serves as gold standard in validation of other methods. Another convolution-based approach was introduced by Bamber [1] and expanded by Meunier [11], in which the image is created by convolution of the imaging system’s point-spread
Data Loading...
