RESUMEN
Introduction: Ultrasound is well-established as an imaging modality for diagnostic and interventional purposes. However, the image quality varies with operator skills as acquiring and interpreting ultrasound images requires extensive training due to the imaging artefacts, the range of acquisition parameters and the variability of patient anatomies. Automating the image acquisition task could improve acquisition reproducibility and quality but training such an algorithm requires large amounts of navigation data, not saved in routine examinations. Methods: We propose a method to generate large amounts of ultrasound images from other modalities and from arbitrary positions, such that this pipeline can later be used by learning algorithms for navigation. We present a novel simulation pipeline which uses segmentations from other modalities, an optimized volumetric data representation and GPU-accelerated Monte Carlo path tracing to generate view-dependent and patient-specific ultrasound images. Results: We extensively validate the correctness of our pipeline with a phantom experiment, where structures' sizes, contrast and speckle noise properties are assessed. Furthermore, we demonstrate its usability to train neural networks for navigation in an echocardiography view classification experiment by generating synthetic images from more than 1,000 patients. Networks pre-trained with our simulations achieve significantly superior performance in settings where large real datasets are not available, especially for under-represented classes. Discussion: The proposed approach allows for fast and accurate patient-specific ultrasound image generation, and its usability for training networks for navigation-related tasks is demonstrated.
RESUMEN
In this work we employ additive manufacturing to print a circular array of micropillars on an aluminium slab turning its top surface into a graded index metasurface for surface acoustic waves (SAW). The graded metasurface reproduces a Luneburg lens capable of focusing plane SAWs to a point. The graded index profile is obtained by exploiting the dispersion properties of the metasurface arising from the well-known resonant coupling between the micropillars (0.5 mm diameter and variable length â¼3 mm) and the surface waves propagating in the substrate. From the analytical formulation of the metasurface's dispersion curves, a slow phase velocity mode is shown to arise from the hybridisation of the surface wave with the pillar resonance. This is used to compute the radial height profile corresponding to the refractive index given by Luneburg's equation. An initial validation of the lens design, achieved through ray theory, shows that ray trajectories have a strong frequency dependence, meaning that the lens will only work on a narrow band. An ultrasonic experiment at 500 kHz where plane SAWs are generated with a piezoelectric transducer and a laser scanner measures the out of plane displacement on the metasurface, validates the actual lens performance and the manufacturing technique. Finally, comparison between the ray analysis and experimental results offers insight into the behaviour of this type of metasurface especially in the proximity of the acoustic bandgaps and highlights the possibility for acoustic shielding.
RESUMEN
Throughout our engineered environment, many materials exhibit a crystalline lattice structure. The orientation of such lattices is crucial in determining functional properties of these structures, including elasticity and magnetism. Hence, tools for determining orientation are highly sought after. Surface acoustic wave velocities in multiple directions can not only highlight the microstructure contrast, but also determine the crystallographic orientation by comparison to a pre-calculated velocity model. This approach has been widely used for the recovery of orientation in cubic materials, with accurate results. However, there is a demand to probe the microstructure in anisotropic crystals - such as hexagonal close packed titanium. Uniquely, hexagonal structure materials exhibit transverse isotropic linear elasticity. In this work, both experimental and simulation results are used to study the discrete effects of both experimental parameters and varying lattice anisotropy across the orientation space, on orientation determination accuracy. Results summarise the theoretical and practical limits of hexagonal orientation determination by linear SAW measurements. Experimental results from a polycrystalline titanium specimen, obtained by electron back scatter diffraction and spatially resolved acoustic spectroscopy show good agreement (errors of Ï1=5.14° and Φ=6.99°). Experimental errors are in accordance with those suggested by simulation, according to the experimental parameters. Further experimental results demonstrate dramatically improved orientation results (Φ error <1°). Demonstrating the possibility of achieving results near the theoretical limit by strict control of the experimental parameters.