Implementação de beamformer adaptativo combinado com generalized sidelobe canceler e filtros para geração de imagens de ultrassom com ondas planas

Abstract

Adaptive beamformer methods have been improving image processing by the Delay and Sum (DAS) method, with increasing resolution and contrast. Even so, the Minimum Variance (MV) method has much to be improved. Thus, new adaptive beamformer techniques have been studied, one of which is Generalized Sidelobe Canceler (GSC) which can considerably improve the lateral resolution of the medical ultrasound images, but the image contrast is not satisfactory. Another technique is the beamformer Eigenspace based on Minimum Variance (EBMV), which is an adaptive beamformer that can suppress much of the noise and provide images with higher contrast and resolution compared to the DAS and the MV methods. Filters can also be applied after image processing to reduce speckle signals present in ultrasound images such as the Wiener, Lee and Kuan. In this work we combined the beamformer adaptive EBMV method with Generalized Sidelobe Canceler and adaptive filters for plane wave ultrasound imaging. All the simulated, in phantom and in vivo data were obtained from Plane Wave Challenge in Medical Ultrasound Imaging (PICMUS) website. The data were acquired using a Verasonics Vantage 128 equipment using a L11-4v 128 elements transducer. The results were compared with DAS and MV. The EBGSC-Kuan method did not show great improvements comparing to the other techniques, whereas the EBGSC-Lee technique, presented good improvements in contrast and reduction of lateral lobes, being an indication to carry out more research with this technique. The EBGSC-Wiener method showed the best improvements in contrast aspects, promoting an improvement of 81% in relation to the DAS method and 67% in relation to the EBGSC method. Also, it presented a reduction of lateral lobe of -59.72 dB and -33.28 dB, compared with DAS and EBGSC techniques, respectively. The EBGSC-Wiener method was the one that best presented Geometric Distortion Ratio (RDG), close to 1, which represents less distortion, being considered a good option to work in the plane wave ultrasound imaging.