Abstract:PurposeUltrasound (US) imaging has been widely used in diagnosis, image‐guided intervention, and therapy, where high‐quality three‐dimensional (3D) images are highly desired from sparsely acquired two‐dimensional (2D) images. This study aims to develop a deep learning‐based algorithm to reconstruct high‐resolution (HR) 3D US images only reliant on the acquired sparsely distributed 2D images.MethodsWe propose a self‐supervised learning framework using cycle‐consistent generative adversarial network (cycleGAN), where two independent cycleGAN models are trained with paired original US images and two sets of low‐resolution (LR) US images, respectively. The two sets of LR US images are obtained through down‐sampling the original US images along the two axes, respectively. In US imaging, in‐plane spatial resolution is generally much higher than through‐plane resolution. By learning the mapping from down‐sampled in‐plane LR images to original HR US images, cycleGAN can generate through‐plane HR images from original sparely distributed 2D images. Finally, HR 3D US images are reconstructed by combining the generated 2D images from the two cycleGAN models.ResultsThe proposed method was assessed on two different datasets. One is automatic breast ultrasound (ABUS) images from 70 breast cancer patients, the other is collected from 45 prostate cancer patients. By applying a spatial resolution enhancement factor of 3 to the breast cases, our proposed method achieved the mean absolute error (MAE) value of 0.90 ± 0.15, the peak signal‐to‐noise ratio (PSNR) value of 37.88 ± 0.88 dB, and the visual information fidelity (VIF) value of 0.69 ± 0.01, which significantly outperforms bicubic interpolation. Similar performances have been achieved using the enhancement factor of 5 in these breast cases and using the enhancement factors of 5 and 10 in the prostate cases.ConclusionsWe have proposed and investigated a new deep learning‐based algorithm for reconstructing HR 3D US images from sparely acquired 2D images. Significant improvement on through‐plane resolution has been achieved by only using the acquired 2D images without any external atlas images. Its self‐supervision capability could accelerate HR US imaging.

Unsupervised Learning for 3D Ultrasonic Data Compression

Data Compression for Ultrasonic Microstructure Scattering Signals using Unsupervised Neural Networks

Machine Learning and Modeling of Ultrasonic Signals for High-Fidelity Data Compression

3-D Ultrasonic Image Reconstruction in Frequency Domain Using a Virtual Transducer Model.

Adaptive quantized PCA with 3D prediction and positive packing for lossless compression of ultraspectral sounder data

Deep data compression for approximate ultrasonic image formation

Spatial-frequency parallel subsampling for distributed compressive sensing in ultrasonic imaging inspection

Stable 3D Deep Convolutional Autoencoder Method for Ultrasonic Testing of Defects in Polymer Composites

Compression of Pulsed Infrared Thermography Data With Unsupervised Learning for Nondestructive Evaluation of Additively Manufactured Metals

Compressed Sensing of Underwater Acoustic Signals Via Structured Approximation $L_0$-Norm

Self-supervised Learning for Accelerated 3D High-resolution Ultrasound Imaging.

Research on Ultrasonic Image Compression in Logging While Drilling: an Asymmetric Convolutional Autoencoder

Deep Compressed Learning for 3D Seismic Inversion

An Underwater Acoustic Data Compression Method Based on Compressed Sensing

3-DUSSS: 3-Dimensional Ultrasonic Self Supervised Segmentation

Three-dimensional Ultrasonic Data Smooth and Endocardial Extraction

Multiresolution Cube Propagation for 3-D Ultrasound Image Reconstruction

Deep Unfolded Robust PCA With Application to Clutter Suppression in Ultrasound

Deep learning-based self-high-resolution for 3D ultrasound imaging

Deep Spectral Methods for Unsupervised Ultrasound Image Interpretation

Efficient Medical Instrument Detection in 3D Volumetric Ultrasound Data