Ultrasonic B-scans Research Articles

Ultrasonic array tomography-oriented subsurface crack recognition and cross-section image reconstruction of reinforced concrete structure using deep neural networks

Manual interpretation from technicians with prior expertise knowledge is still needed for ultrasonic array tomography when detecting subsurface damages in reinforced concrete (RC) structures. Ultrasonic scans were usually utilized in qualitative manners that have not been leveraged for quantification purposes in prior studies and practical engineering. This study presented an ultrasonic array tomography-oriented subsurface crack recognition and pixel-wise cross-section image reconstruction method for RC structures based on deep neural networks (DNN). B-scans were acquired by testing entity RC components with preset artificial cracks using shear-wave low-frequency ultrasonic array. A DNN with the basic encoder-decoder architecture was developed, which was improved by introducing skip connections and residual modules to adapt to semantic structure of ultrasonic B-scans. A supervised generative adversarial DNN was proposed to generate fake but convincing B-scans for data augmentation. Results reported that F-scores were higher than 78 %, and satisfactory agreements were observed between ground truth annotations and reconstructed cross-section images, substantiating the effectiveness of the proposed method. Furthermore, with the physical information-assisted registration, individual local reconstructed cross-section images were combined to global representations indicating crack distribution in the entire component to facilitate convincing structural assessments.

Generative adversarial network with object detector discriminator for enhanced defect detection on ultrasonic B-scans

Non-destructive testing is a set of techniques for defect detection in materials. While the set of imaging techniques is manifold, ultrasonic imaging is the one used the most. The analysis is mainly performed by human inspectors manually analyzing the acquired images. A low number of defects in real ultrasonic inspections and legal issues concerning data from such inspections make it difficult to obtain proper results from automatic ultrasonic image (B-scan) analysis. The goal of presented research is to obtain an improvement of the detection results by expanding the training data set with realistic synthetic samples. In this paper, we present a novel deep learning Generative Adversarial Network model for generating realistic ultrasonic B-scans with defects in distinct locations. Furthermore, we show that generated B-scans can be used for synthetic data augmentation, and can improve the performances of deep convolutional neural object detection networks. Our novel method was developed on a dataset with almost 4000 images and more than 6000 annotated defects. When trained only on real data, detector can achieve an average precision of 70%. By training only on generated data the results increased to 72%, and by mixing generated and real data we achieve almost 76% average precision. We believe that synthetic data generation can generalize to other tasks with limited data. It could also be used for training human personnel.

Baseline-free defect evaluation of complex-microstructure composites using frequency-dependent ultrasound reflections

In this study, baseline-free defect characterization for both flat and ply-drop composites were proposed using frequency dependence of ultrasound reflections. This frequency dependence of ultrasound interaction with delamination and resin layers in composites were investigated through CIVA® simulation and experimental investigation. To extract quantitative data of frequency dependence of ultrasonic reflections in composites with a single test, synchrosqueezed wavelet transform was adopted to decompose the acquired wide-band raw signals into narrow-band frequencies. Quantitative analysis concluded that dependence of ultrasound reflections from defects on frequency in composites are affected by wavelength-to-defect ratio, polymer viscoelasticity, frequency-dependent inter-ply reflections and waviness-induced scattering. Considering these factors, our study found that volumetric defect reflections have a unique frequency dependence. This allows volumetric defect identification and characterization in ultrasonic B-scans to be achieved without comparison to a baseline reference. The proposed method shows a promising solution for non-destructive evaluation of multi-layer composites with diverse microstructures.

Automated and antidispersive coherent and incoherent noise reduction of waveforms that contain a reference pulse

Fully automated and antidispersive noise reduction of a waveform containing a reference pulse is often necessary for more efficient, accurate, and reliable interpretation of signals – such as highly attenuated waves in ultrasonic NDE. We propose two methods, pulse specification (PS) and coherent enhanced pulse specification (CEPS), for improving the signal-to-noise ratio (SNR) of coherent pulse amplitudes against both incoherent and coherent noise, as well as apply complementary antidispersive pulse generation (APG) to localise the pulse upon reception. We have demonstrated PS and CEPS effectiveness, with and without APG, using two examples of ultrasonic inspection, by comparison with established methods and demonstrated improvements in SNR up to a factor of 2 (6.02 dB) and never less than 1 (0 dB) for favoured pulse specifications. This has benefits in applications such as NDE, medical ultrasonics, seismology, and acoustics.We deconvolve waveforms using a reference pulse time windowed from said waveforms. These waveforms are convolved with pulses that have centre frequencies and bandwidths that are optimally and automatically specifiable for that unprocessed waveform. We combine independent measurements of the waveforms using PS and CEPS to maximise the coherent pulse amplitude of known scatterers and minimise the noise content. CEPS also enhances the coherent scatterer amplitude to significantly greater than achieved using an established filtered time shift method – where the time shift is subjective. Unlike such time shift methods, PS and CEPS do not reduce the amplitude of the coherent pulse signals when reducing noise.Automated and antidispersive coherent and incoherent noise reduction by pulse specification has been demonstrated to be effective in ultrasonic NDE of both damping and scattering media. Suggestions for further applications of this method have already included application to ultrasonic NDE of composite media, application to ultrasonic guided wave NDE, ultrasonic NDE imaging that can include ultrasonic B-scans, and more. This is greatly encouraging regarding the potential immediate applicability of noise reduction by pulse specification.

Ultrasound Image Intensity Nonuniformity Correction by Combining Intensity and Spatial Information

Medical ultrasonic B-scans often suffer from intensity inhomogeneities that originates from the nonuniform attenuation properties of the sonic beam within the body. In order to correct signal attenuation in the tissue, time gain compensation (TGC) is routinely used in medical ultrasound scanners. However, TGC assumes a uniform attenuation coefficient for all body tissues. Since this assumption is evidently inaccurate, over-amplification or under-amplification sometimes appear. This is a major problem for intensity-based, automatic segmentation of video-intensity images because conventional threshold-based or intensity-statistic-based approaches do not work well in the presence of such image distortions. The main contribution of this paper is that additional spatial image features are incorporated to improve inhomogeneity correction and to make it more dynamic besides most commonly used intensity features, so that local intensity variations can be corrected more efficiently. The degraded image is corrected by the inverse of the image degradation model. The image degradation process is described by a linear model, consisting of a multiplicative and an additive component which are modeled by a combination of smoothly varying basis functions. Spatial information about intensity nonuniformity is obtained using cubic spline smoothing and entropy minimizing. Gray-level histogram information of the image corrupted by intensity inhomogeneity is exploited from a signal processing perspective. We explain how this model can be related to the ultrasonic physics of image formation to justify our approach. Experiments are presented on synthetic images and real US data to evaluate quantitatively the accuracy of the method.

A statistical model for the ultrasonic backscattered echo from tissue containing microcalcifications

Microcalcifications, generally seen in B-mode breast images as bright spots, often point to the presence of malignancies. Currently, statistical models to describe the signal strengths from tissue containing microcalcifications rely on the Rician or Nakagami densities. Because the Nakagami density requires the existence of randomly located scatterers with random cross sections, microcalcifications manifesting as isolated strong scatterers in the scattering volume makes it less suitable. Additionally, the absence of any regularly spaced scatterers also makes the Rician density less suitable. In this work, it is proposed that a 3-parameter McKay density with heavier tails than the Nakagami density, and hence more severe intensity fluctuations (speckle), might be a better fit to describe echo from tissue containing microcalcifications. This density is developed using a physical description of scattering and its characteristics are explored through random number simulation. Results demonstrate that the McKay densities have higher levels of speckle quantified through the speckle factor than the corresponding gamma densities (Nakagami for the envelope). Some preliminary analysis of tissue-mimicking phantoms modified to include microcalcification-like regions shows that higher speckle factors (seen with the McKay density) can be used to isolate and display microcalcifications. It is expected that this novel approach based on the McKay density will lead to newer and simpler means to detect and identify microcalcifications in ultrasonic B-scans.

Strain imaging of corneal tissue with an ultrasound elasticity microscope.

We hypothesize that high-resolution elasticity measurements can guide corrective refractive surgery of the cornea. Elasticity measurements would improve surgical outcomes by adding biomechanical information not used in existing clinical nomograms. As an initial investigation, we determined the usefulness and evaluated the ability of our ultrasound elasticity microscope by measuring strain ex vivo in an intact porcine eye globe. Strain was predicted with a finite element model guided by direct mechanical measurements of corneal elasticity. Next, a porcine cornea was deformed with a slitted plate while being imaged with ultrasound. For high spatial resolution, the ultrasound elasticity microscope uses a 50 MHz transducer with a 1.4 f/number. It produces high-quality conventional ultrasonic B-scans over large thicknesses by confocal processing. Strain was calculated from tracking speckle in these images after deformation. This technique is compatible with in vivo measurements. Compressional and expansional deformations were the same order of magnitude from -3.5% to as great as +3.5%. Strain imaging indicated the stroma expanded into the slit of the deformation plate while Bowman's layer compressed. This bipolar variation within a specimen is unusual. Within the stroma, a variation of strain with depth was measured suggesting a distribution of elasticity. Results compared favorably with the finite element model. An ultrasound elasticity microscope can produce high-resolution strain images throughout the corneal depth. Various layers with different elastic properties appeared as different strains in the images.

Characterization of ultrasonic B-scans using non-rayleigh statistics

The envelope of the ultrasonic echo backscattered from tissues is modeled using non-Rayleigh statistics, namely the K-distribution. We show that the K-distribution can be used to describe the envelope of the echo, and its parameters may be used to distinguish between different regions in ultrasonic B-scan images. The validity of the model is tested using phantoms. Preliminary results indicate that the parameters of the K-distribution may be used to separate targets, in which the number of scatterers or the scattering cross-sections differ from that of the background. The method of employing the parameters of the K-distribution appears to be better than the methods based solely on the signal-to-noise ratio.

Use of non-Rayleigh statistics for the identification of tumors in ultrasonic B-scans of the breast

A model for the scattering of ultrasound from breast tissue is proposed. The model is based on the use of non-Rayleigh statistics, specifically the K distribution to describe the backscattered echo from the tissue. A multiparameter test based on this model has been designed to characterize the tissue. The data from the B-scan images of the breasts of 6 different patients were analyzed using this model. The results indicate that the non-Rayleigh statistics seem to be useful in characterizing and identifying malignant, benign, and normal tissue regions.

Ultrasonic B‐scans of the facial/oral periodontium in pigs

In vitro ultrasonic scans of the oral/facial periodontium of pigs were obtained using a special experimental design and with the help of a newly developed ultrasonic measuring and analysis system. These two-dimensional images distinguish themselves from the one-dimensional RF-echograms and A-scans in that they enable the viewer to have a quick overview of the relationship between different interfaces in a scanned sector. A selection of ultrasonic B-scans proves that the scanned tissues such as gingiva, bone, periodontal ligament, tooth surface and even structures of the inner part of the tooth may be clearly imaged and distinctly differentiated. Furthermore, we demonstrate with manipulations of the pigs periodontium, that the interpretation of the pictures is correct. The primary goal of our investigation was to assess whether the ultrasonic procedure is suitable for the determination of the height of the facial/oral bone crest. On the basis of this study, we may confirm this supposition for the imaging in ultrasonic B-scans.

Response to comment on Vaknine, R. and Lorenz, W.J. Lateral filtering of medical ultrasonic B-scans before image generation, [formula omitted], 152–158 (1984)

Response to comment on Vaknine, R. and Lorenz, W.J. Lateral filtering of medical ultrasonic B-scans before image generation, [formula omitted], 152–158 (1984)

Lateral filtering of medical ultrasonic B-scans before image generation

A method is presented for improving the lateral resolution of diagnostic B-scan images. Wiener filtering is applied on sequences of A-scan echo amplitudes acquired during B-scanning, using a simple scanning action. The first applications to clinical B-scan data show an improvement in the image quality. The filtered images obtained by this technique may provide an improvement in the performance of pattern recognition methods on medical ultrasonic image data.

Rician statistics and signal detectability in ultrasonic B-scans

Rician statistics and signal detectability in ultrasonic B-scans

Digital ultrasonic imaging with microprocessor manipulation.

The author describes the development of a digital scan converter system which accepts ultrasonic B-scans in "oscilloscope" format and displays them in video format. The system involves the use of a microprocessor, the function of which is outlined in the paper. A brief comparison with a system employing a scan converter tube is given.

Ultrasonic B-Scanning for radiation therapy planning and follow-up of superficial tumors

Ultrasonic B-scans are employed for delineating the boundaries of superficial, soft-tissue tumors for radiation therapy treatment planning. Full-scale anatomical information derived from magnified echograms is used for obtaining optimal photon beam dose distributions with the aid of a mini-computer. The feasibility of using ultrasound scans for assisting with electron beam radiation therapy planning is demonstrated. Tumor response following initiation of therapy may be evaluated using echograms for determining overall tumor dimensions or for computing the tumor volume. The use of ultrasonic B-scans for developing a 3-dimensional isometric display of tumor contours is discussed.

Quantitation of tumor volumes and response to therapy with ultrasound B-scans.

A technique employing serial ultrasonic B-scans and a computer for measuring volumes of internal anatomical structures is described. Phantom studies indicate that this method is accurate to within 5% for phantom volumes in the 200–2000 cm3 range. The technique was applied to monitor the spleen-tumor volume in a patient with chronic lymphocytic leukemia and the liver-tumor volume in a patient with retroperitoneal rhabdomyosarcoma. The data suggest that the method is clinically useful for determining normal organ and tumor volumes and for evaluating response to therapy.

Ultrasonic B-scanning of the heart with an ECG gated television display

A method is described for obtaining ultrasonic B-scans of the heart at various phases of the cardiac cycle. Preliminary results indicate that this technique can be a valuable complement to time-position scanning of the heart and can provide useful information about the cardiac geometry. The system described can be used with any ultrasonic B-scan machine.

B-scan ultrasonography of orbital lymphangiomas.

Unilateral exophthalmos is an important clinical ophthalmic problem. A clinical diagnosis of the cause of the exophthalmos is often very difficult. B-scan ultrasonography is an important new diagnostic modality with great potential in determination of the nature of orbital pathology. The earliest ophthalmic applications of ultrasound were reported by Mundt and Hughes (I956). A-scan ultrasound, as described by Oksala and Lehtinen (I957), Gernet (I963), and Ossoinig (I969), is now widely used. B-scan ultrasound was initially utilized in only a few laboratories (Baum and Greenwood, I958; Purnell, I965) because of the high cost and complexity of early B-scan instrumentation. B-scan equipment has recently become more readily available (Coleman, Konig, and Katz, I 969). B-scan ultrasound yields two-dimensional acoustic sections of the eye and orbit and is the technique of choice for orbital diagnosis. The difference between ultrasonic Aand B-scans is shown diagrammatically in Fig. I.

Multiplexing ultrasonic wave fronts by holoraphy.

Easily viewable, three-dimensional images have been produced from information derived from the passage of sound through the head of a living human subject. In this technique a new form of holographic multiplexing is used to construct the three-dimensional image from two-dimensional ultrasonic B-scans taken in many separated planes.