Research Projects

Machine Learning-Based B-mode Imaging

Ultrasound B-mode images display the “echogenicity”, or acoustic brightness, of the underlying soft tissue. Traditional B-mode images are formed using delay-and-sum (DAS) beamforming, which displays the magnitude of the mean complex channel signal. DAS is a robust technique that is used ubiquitously in medical ultrasound imaging systems, but is fundamentally subject to noise artifacts such as speckle. Machine learning and deep learning techniques present an alternative way to tackle echogenicity estimation. Rather than using a deterministic algorithm such as DAS, neural networks are trained empirically to reconstruct optimal B-mode images as quantified by metrics such as mean absolute error, mean squared error, structural similarity, etc. We have recently demonstrated that simple fully convolutional neural networks produce more accurate echogenicity estimates than DAS in ultrasound simulations.

Fig. 1. In vivo images were acquired in the liver (left two images) and kidney (right two images) of a healthy volunteer. In each pair of images, the middle sectors were reconstructed using (left) DAS and (right) the simulation-trained neural network (NN). The NN reduces speckle while preserving fine structures.

Fig. 2. In vivo images were acquired in the liver of a volunteer with a focal lesion surrounded by a fluid capsule. As seen in the zoomed inset, the NN substantially reduces speckle while preserving many of the sharp features present within the DAS image.

An open source implementation of neural network-based B-mode imaging is provided here:

https://gitlab.com/dongwoon.hyun/nn_bmode

Training, validation, and testing data are also provided:

https://drive.google.com/drive/folders/1cNUsUhJs4KM_ujxl_Vs1Hl9vDJuUgS3K?usp=sharing

 

This code is free to use and is covered by the Apache v2 license. Please cite the following reference when using this code or data:

D. Hyun, L. L. Brickson, K. T. Looby, and J. J. Dahl. "Beamforming and Speckle Reduction Using Neural Networks." IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, 66(5), pp. 898-910, May 2019. doi: 10.1109/TUFFC.2019.2903795