Visual geometry Group-UNet: Deep learning ultrasonic image reconstruction for curved parts

Ultrasound image super-resolution reconstruction based on semi-supervised CycleGAN

In ultrasonic testing, diffraction artifacts generated around defects increase the challenge of quantitatively characterizing defects. In this paper, we propose a label-enhanced semi-supervised CycleGAN network model, referred to as LESS-CycleGAN, which is a conditional cycle generative adversarial network designed for accurately characterizing defect morphology in ultrasonic testing images. The proposed method introduces paired cross-domain image samples during model training to achieve a defect transformation between the ultrasound image domain and the morphology image domain, thereby eliminating artifacts. Furthermore, the method incorporates a novel authenticity loss function to ensure high-precision defect reconstruction capability. To validate the effectiveness and robustness of the model, we use simulated 2D images of defects and corresponding ultrasonic detection images as training and test sets, and an actual ultrasonic phased array image of a test block as the validation set to evaluate the model's application performance. The experimental results demonstrate that the proposed method is convenient and effective, achieving subwavelength-scale defect reconstruction with good robustness.

Artificial intelligence and duplex ultrasound for detection of carotid artery disease

OBJECTIVE

The aim of this study is to evaluate the feasibility, applicability and accuracy of artificial intelligence (AI) in the detection of normal versus carotid artery disease through greyscale static duplex ultrasound (DUS) images.

METHODS

A prospective image acquisition of individuals undergoing duplex sonography for the suspicion of carotid artery disease at a single hospital was conducted. A total of n = 156 images of normal and stenotic carotid arteries (based on NASCET criteria) were evaluated by using geometry group network based on convolutional neural network (CNN) architecture. Outcome was reported based on sensitivity, specificity and accuracy of the network (artificial intelligence) for detecting normal versus stenotic carotid arteries as well as various categories of carotid artery stenosis.

RESULTS

The overall sensitivity, specificity and accuracy of AI in the detection of normal carotid artery was 91%, 86% and 92%, respectively, and for any carotid artery stenosis was 87%, 82% and 90%, respectively. Subgroup analyses demonstrated that the network has the ability to detect stenotic carotid artery images (<50%) versus normal with a sensitivity of 92%, specificity of 87% and an accuracy of 94%. This value (sensitivity, specificity and accuracy) for group of 50-75% stenosis versus normal was 84%, 80% and 88% and for carotid artery disease of more than 75% was 90%, 83% and 92%, respectively.

CONCLUSION

This study demonstrates the feasibility, applicability and accuracy of artificial intelligence in the detection of carotid artery disease in greyscale static DUS images. This network has the potential to be used as a stand-alone software or to be embedded in any DUS machine. This can enhance carotid artery disease recognition with limited or no vascular experience or serve as a stratification tool for tertiary referral, further imaging and overall management.

A multiscale residual U-net architecture for super-resolution ultrasonic phased array imaging from full matrix capture data

Ultrasonic phased array imaging using full-matrix capture (FMC) has raised great interest among various communities, including the nondestructive testing community, as it makes full use of the echo space to provide preferable visualization performance of inhomogeneities. The conventional way of FMC data postprocessing for imaging is through beamforming approaches, such as delay-and-sum, which suffers from limited imaging resolution and contrast-to-noise ratio. To tackle these difficulties, we propose a deep learning (DL)-based image forming approach, termed FMC-Net, to reconstruct high-quality ultrasonic images directly from FMC data. Benefitting from the remarkable capability of DL to approximate nonlinear mapping, the developed FMC-Net automatically models the underlying nonlinear wave-matter interactions; thus, it is trained end-to-end to link the FMC data to the spatial distribution of the acoustic scattering coefficient of the inspected object. Specifically, the FMC-Net is an encoder-decoder architecture composed of multiscale residual modules that make local perception at different scales for the transmitter-receiver pair combinations in the FMC data. We numerically and experimentally compared the DL imaging results to the total focusing method and wavenumber algorithm and demonstrated that the proposed FMC-Net remarkably outperforms conventional methods in terms of exceeding resolution limit and visualizing subwavelength defects. It is expected that the proposed DL approach can benefit a variety of ultrasonic array imaging applications.

Ultrasonic tomography imaging enhancement approach based on deep convolutional neural networks

Abstract

The containment liner plate (CLP) is a thin layer of carbon steel material applied as a base for concrete structures protecting nuclear material. The structural health monitoring of the CLP is critical to ensure the safety of nuclear power plants. Hidden defects in the CLP can be identified utilizing ultrasonic tomographic imaging techniques such as the reconstruction algorithm for the probabilistic inspection of damage (RAPID) methodology. However, Lamb waves have a multimodal dispersion feature, which makes the selection of a single mode more difficult. Thus, sensitivity analysis was utilized since it allows for the determination of each mode's level of sensitivity as a function of frequency; the S0 mode was chosen after examining the sensitivity. Even though proper Lamb wave mode was selected, the tomographic image had blurred zones. Blurring reduces the precision of an ultrasonic image and makes it more difficult to distinguish the dimensions of the flaw. To enhance the tomographic image of the CLP, deep learning architecture such as U-Net was utilized for the segmentation of the experimental ultrasonic tomographic image, which includes an encoder and decoder part for better visualization of the tomographic image. Nevertheless, collecting enough ultrasonic images to train the U-Net model was not economically feasible, and only a small number of the CLP specimens can be tested. Thus, it was necessary to utilize transfer learning and get the values of the parameters from a pre-trained model with a much larger dataset as a starting point for a new task, rather than training a new model from scratch. Through these deep learning approaches, we were able to eliminate the blurred section of the ultrasonic tomography, leading to images with clear edges of defects and no blurred zones.

Graphical abstract

A review of synthetic and augmented training data for machine learning in ultrasonic non-destructive evaluation

Ultrasonic Testing (UT) has seen increasing application of machine learning (ML) in recent years, promoting higher-level automation and decision-making in flaw detection and classification. Building a generalized training dataset to apply ML in non-destructive evaluation (NDE), and thus UT, is exceptionally difficult since data on pristine and representative flawed specimens are needed. Yet, in most UT test cases flawed specimen data is inherently rare making data coverage the leading problem when applying ML. Common data augmentation (DA) strategies offer limited solutions as they don't increase the dataset variance, which can lead to overfitting of the training data. The virtual defect method and the recent application of generative adversarial neural networks (GANs) in UT are sophisticated DA methods targeting to solve this problem. On the other hand, well-established research in modeling ultrasonic wave propagations allows for the generation of synthetic UT training data. In this context, we present a first thematic review to summarize the progress of the last decades on synthetic and augmented UT training data in NDE. Additionally, an overview of methods for synthetic UT data generation and augmentation is presented. Among numerical methods such as finite element, finite difference, and elastodynamic finite integration methods, semi-analytical methods such as general point source synthesis, superposition of Gaussian beams, and the pencil method as well as other UT modeling software are presented and discussed. Likewise, existing DA methods for one- and multidimensional UT data, feature space augmentation, and GANs for augmentation are presented and discussed. The paper closes with an in-detail discussion of the advantages and limitations of existing methods for both synthetic UT training data generation and DA of UT data to aid the decision-making of the reader for the application to specific test cases.

Laser ultrasonic imaging of complex defects with full-matrix capture and deep-learning extraction

Phased array-based full-matrix ultrasonic imaging has been the golden standard for the non-destructive evaluation of critical components. However, the piezoelectric phased array cannot be applied in hazardous environments and online monitoring due to its couplant requirement. The laser ultrasonic technique can readily address these challenging tasks via fully non-contact inspection, but low detection sensitivity and complicated wave mode conversion hamper its practical applications. The laser-induced full-matrix ultrasonic imaging of complex defects was displayed in this study. Full matrix data acquisition and deep learning method were adapted to the laser ultrasonic technique to overcome the existing challenges. For proof-of-concept demonstrations, simulations and experiments were conducted on an aluminum sample with representative defects. Numerical and experimental results showed good agreement, revealing the excellent imaging performance of proposed method. Compared with the total focusing method based on ray-trace model, the deep learning method could create superior images with additional quantitative information through end-to-end networks, which use the hierarchical features and generate details from all the relevant imaging and physical characteristics information. The proposed method may help assess defect formation and development at the early stage in a hazardous environment and understand the in-situ manufacturing process due to its couplant-free nature.

VGG-UNet/VGG-SegNet Supported Automatic Segmentation of Endoplasmic Reticulum Network in Fluorescence Microscopy Images

This research work aims to implement an automated segmentation process to extract the endoplasmic reticulum (ER) network in fluorescence microscopy images (FMI) using pretrained convolutional neural network (CNN). The threshold level of the raw FMT is complex, and extraction of the ER network is a challenging task. Hence, an image conversion procedure is initially employed to reduce its complexity. This work employed the pretrained CNN schemes, such as VGG-UNet and VGG-SegNet, to mine the ER network from the chosen FMI test images. The proposed ER segmentation pipeline consists of the following phases; (i) clinical image collection, 16-bit to 8-bit conversion and resizing; (ii) implementation of pretrained VGG-UNet and VGG-SegNet; (iii) extraction of the binary form of ER network; (iv) comparing the mined ER with ground-truth; and (v) computation of image measures and validation. The considered FMI dataset consists of 223 test images, and image augmentation is then implemented to increase these images. The result of this scheme is then confirmed against other CNN methods, such as U-Net, SegNet, and Res-UNet. The experimental outcome confirms a segmentation accuracy of >98% with VGG-UNet and VGG-SegNet. The results of this research authenticate that the proposed pipeline can be considered to examine the clinical-grade FMI.

Perioperative Nursing Management of Patients Undergoing Laparoscopic Ovarian Cystectomy Guided by Ultrasound Imaging under Intelligent Algorithm

This study was aimed at exploring the application value of ultrasonic imaging-guided laparoscopic ovarian cystectomy after denoising by intelligent algorithms in perioperative nursing intervention of patients. In this study, convolutional downsampling was introduced to the UNet model, based on which the residual structure and Recon module were added to improve the UNet denoising model, which was applied to 100 patients who underwent ultrasound imaging-guided laparoscopic ovarian cystectomy. The patients were grouped into a control group receiving conventional nursing and an experimental group receiving perioperative nursing management. The various experimental indicators were comprehensively evaluated. The results revealed that after denoising using the improved UNet model, the ultrasound image showed no unnecessary interference noise, and the image clarity was significantly improved. In the experimental group, the operation time was 55.45 ± 6.13 days, the intraoperative blood loss was 71.52 ± 9.87 days, the postoperative exhaust time was 1.9 ± 0.73 days, the time to get out of bed was 1.2 ± 0.85 days, the complication rate was 8%, the hospitalization time was 7.3 ± 2.6 days, and the nursing satisfaction rate reached 98%. All above aspects were significantly better than those of the control group, and the differences were statistically significant (P < 0.05). In short, the improved UNet denoising model can effectively eliminate the interference noise in ultrasound and restore high-quality ultrasound images. Perioperative nursing intervention can accelerate the recovery speed of patients, reduce the complication rate, and shorten the length of stay in hospital. Therefore, it was worthy of being widely used in clinical nursing.

Introduction to the special issue on machine learning in acoustics

The use of machine learning (ML) in acoustics has received much attention in the last decade. ML is unique in that it can be applied to all areas of acoustics. ML has transformative potentials as it can extract statistically based new information about events observed in acoustic data. Acoustic data provide scientific and engineering insight ranging from biology and communications to ocean and Earth science. This special issue included 61 papers, illustrating the very diverse applications of ML in acoustics.