Three-dimensional Volume Data Research Articles

An effective ultrasound fetal palate screening software based on the “sequential sector scan through the oral fissure” and three-dimensional ultrasound

BackgroundOrofacial clefts are one of the most common congenital malformations of the fetal face and ultrasound is mainly responsible for its diagnosis. It is difficult to view the fetal palate, so there is currently no unified standard for fetal palate screening, and the diagnosis of cleft palate is not included in the relevant prenatal ultrasound screening guidelines. Many prenatal diagnoses for cleft palate are missed due to the lack of effective screening methods. Therefore, it is imperative to increase the display rate of the fetal palate, which would improve the detection rate and diagnostic accuracy for cleft palate. We aim to introduce a fetal palate screening software based on the “sequential sector scan though the oral fissure”, an effective method for fetal palate screening which was verified by our follow up results and three-dimensional ultrasound and to evaluate its feasibility and clinical practicability.MethodsA software was designed and programmed based on “sequential sector scan through the oral fissure” and three-dimensional ultrasound. The three-dimensional ultrasound volume data of the fetal face were imported into the software. Then, the median sagittal plane was taken as the reference interface, the anterior upper margin of the mandibular alveolar bone was selected as the fulcrum, the interval angles, and the number of layers of the sector scan were set, after which the automatic scan was performed. Thus, the sector scan sequential planes of the mandibular alveolar bone, pharynx, soft palate, hard palate, and maxillary alveolar bone were obtained in sequence to display and evaluate the palate. In addition, the feasibility and accuracy of the software in fetal palate displaying and screening was evaluated by actual clinical cases.ResultsFull views of the normal fetal palates and the defective parts of the cleft palates were displayed, and relatively clear sequential tomographic images and continuous dynamic videos were formed after the three-dimensional volume data of 10 normal fetal palates and 10 cleft palates were imported into the software.ConclusionsThe software can display fetal palates more directly which might allow for a new method of fetal palate screening and cleft palate diagnosis.

Convolutional neural network for assisting accuracy of personalized clavicle bone implant designs

<span lang="EN-US">The clavicle is a long bone that tends to be frequently fractured in the midshaft region. The plate and screw fixing method is mainly applied to address this issue. This study aims to construct a clavicle bone implant design with a consideration to achieve a high accuracy and high-quality surface between the plate and the clavicle surface. The computational tomography scanning (CT-scan) image series data were processed using a convolutional neural network (CNN) to classify the clavicle image. The CNN outcomes were gathered as three-dimensional (3D) volume data of clavicle bone. This 3D model was then proposed for the plate design. The CNN testing results of 97.4% for the image clavicle bones classification, whereas the prints of the 3D model from clavicle bone and its plate and screw design reveal compatibility between the bone surface and the plate surface. Overall, the CNN application to the series of CT images could ease the classification of clavicle bone images that would precisely construct the 3D model of clavicle bone and its suitable clavicle bone plate design. This study could contribute as a guideline for other bone plate areas that need to fit the patient’s bone geometry.</span>

Utility of Virtual Bronchoscopy in Tracheal Tumors.

Tracheal tumors are rare neoplasms, most of which are malignant. Conventional bronchoscopy (CB) and biopsy are used in the primary evaluation of these tumors, while Contrast-Enhanced Computed Tomography (CECT) is usually used for assessing extraluminal extension and nodal spread. Virtual bronchoscopy, created by postprocessing the three-dimensional volume data acquired from CECT, attempts to recreate the endoluminal features as seen in CB. The complexity of tracheal surgery necessitates accurate preoperative localization of the lesion and planning. It can assist in evaluation of patients in whom CB cannot be performed and for airway assessment in a preoperative setting. Although it cannot replace CB, it can be resorted to, in specific clinical scenarios. In this pictorial essay, we aim to highlight the technique, role, and clinical utility of Virtual Bronchoscopy (VB) in tracheal tumors.

Sex- and Age-Dependent Wide-Field Choroidal Thickness Differences in Healthy Eyes

In this study, we aimed to map and characterize the choroidal thickness over a wide area from the posterior pole to the vortex vein in normal eyes. This observational study included 146 healthy eyes (63 male). Three-dimensional volume data were acquired to create a choroidal thickness map using swept-source optical coherence tomography. The map was classified as type A if an area with a choroidal thickness >250 µm in the vertical direction from the optic disc, and the area corresponding to the watershed was not observed, or as type B if such an area was observed. The relationship between the ratio of groups A to B and age was compared by classifying the age for three age groups: <40, 40-60, and >60 years in men and women. In men and women, 69.8% and 49.4% were classified as type A, respectively, with significant sex differences (p = 0.013). The proportion of type B decreased with increasing age in both the sexes. There was a significant difference between ≤60 and >60 years in men and between ≤40 and >40 years in women (p < 0.05). To conclude, the wide-area choroidal thickness and the age-dependent changes in healthy eyes differed between the sexes.

Evaluation of Scapholunate Injury and Repair with Dynamic (4D) CT: A Preliminary Report of Two Cases.

Background In predynamic or dynamic scapholunate (SL) instability, standard diagnostic imaging may not identify SL interosseous ligament (SLIL) injury, leading to delayed detection and intervention. This study describes the use of four-dimensional computed tomography (4DCT) in identifying early SLIL injury and following injured wrists to 1-year postoperatively. Description of Technique 4DCT acquires a series of three-dimensional volume data with high temporal resolution (66 ms). 4DCT-derived arthrokinematic data can be used as biomarkers of ligament integrity. Patients and Methods This study presents the use of 4DCT in a two-participant case series to assess changes in arthrokinematics following unilateral SLIL injury preoperatively and 1-year postoperatively. Patients were treated with volar ligament repair with volar capsulodesis and arthroscopic dorsal capsulodesis. Arthrokinematics were compared between uninjured, preoperative injured, and postoperative injured (repaired) wrists. Results 4DCT detected changes in interosseous distances during flexion-extension and radioulnar deviation. Generally, radioscaphoid joint distances were greatest in the uninjured wrist during flexion-extension and radioulnar deviation, and SL interval distances were smallest in the uninjured wrist during flexion-extension and radioulnar deviation. Conclusion 4DCT provides insight into carpal arthrokinematics during motion. Distances between the radioscaphoid joint and SL interval can be displayed as proximity maps or as simplified descriptive statistics to facilitate comparisons between wrists and time points. These data offer insight into areas of concern for decreased interosseous distance and increased intercarpal diastasis. This method may allow surgeons to assess whether (1) injury can be visualized during motion, (2) surgery repaired the injury, and (3) surgery restored normal carpal motion. Level of Evidence Level IV, Case series.

A new method for identifying micro fractures and characterizing fractures of different scales

Oil and gas exploration professionals have begun to focus more on unconventional oil and gas reserves in recent years as a result of their increased efforts. Fractures have a significant impact on the permeability and connectivity of reservoirs as a crucial component of rock mechanics and hydraulics, which directly affects the production of oil and gas. The identification of fracture development zones or micro faults, as well as how to adequately define the fracturing model, have thus become crucial and pressing issues in the forecast of oil and gas reservoirs. In this study, we decompose the three-dimensional seismic data volume in a site in order to obtain the single frequency data volume that can be calculated using the ant tracking technique. We do this by taking advantage of the synchronous extrusion improvement of short time Fourier transform in time-frequency focusing. Coupled with the advanced DFN model, the extracted data are calibrated in various rock attributes to restore the morphology and characteristics of fractures. The findings demonstrate that this method is capable of providing not only a precise outline of micro fractures but also a reflection of the characteristics of fractures at various scales, including structure and associated properties. The precision and applicability of this method are confirmed in this paper, which is significant as a reference for the oil and gas exploration industry.

Three-dimensional reconstructing undersampled photoacoustic microscopy images using deep learning

Spatial sampling density and data size are important determinants of the imaging speed of photoacoustic microscopy (PAM). Therefore, undersampling methods that reduce the number of scanning points are typically adopted to enhance the imaging speed of PAM by increasing the scanning step size. Since undersampling methods sacrifice spatial sampling density, by considering the number of data points, data size, and the characteristics of PAM that provides three-dimensional (3D) volume data, in this study, we newly reported deep learning-based fully reconstructing the undersampled 3D PAM data. The results of quantitative analyses demonstrate that the proposed method exhibits robustness and outperforms interpolation-based reconstruction methods at various undersampling ratios, enhancing the PAM system performance with 80-times faster-imaging speed and 800-times lower data size. The proposed method is demonstrated to be the closest model that can be used under experimental conditions, effectively shortening the imaging time with significantly reduced data size for processing.

Could premature ventricular contractions lead to atrial remodeling?

Premature ventricular contraction (PVC) is a frequent kind of arrhythmia that affects around 1% of the general population. While PVC most frequently impairs ventricular function in structurally normal heart, retrograde ventriculo-atrial conduction can occur in people with PVC. These retrograde atrial activations may mimic pulmonary vein-derived atrial ectopies. As a result, PVC may raise the risk of atrial fibrillation (AF) by retrograde ventriculo-atrial conduction. The Four-Dimensional Automated Left Atrial Quantification (4D Auto LAQ) tool is a left atrial analytical approach that utilizes three-dimensional volume data to quantify the volume, as well as LA longitudinal and circumferential strains. The purpose of this study was to determine if clinical diagnosis of PVC is connected with abnormal LA function as determined by LA strain evaluation utilizing a 4D Auto LAQ compared to the healthy population. The 58 patients with frequent PVCs and 53 healthy volunteers as a control group were enrolled in the study. Imaging was performed using the GE Vivid E95 echocardiography equipment (GE Healthcare; Vingmed Ultrasound, Horten, Norway) equipped with an M5S probe (frequency range: 1.5-4.6 MHz) and a 4V probe (frequency range: 1.5-4.0 MHz). Images were imported into and were selected for analysis using the EchoPAC203 software (GE Healthcare). The analysis mode was selected, followed by the volume and 4D Auto LAQ submodes. Following that, the sample point was positioned in the center of the mitral orifice in each of the three planes. The review function was used to acquire the LA parameters measured by 4D Auto LAQ, including volume and strain parameters. The maximum left atrial volume (LAVmax) and minimal left atrial volume (LAVmin) were significantly higher in the patient group (38.91 ± 9.72 vs. 46.31 ± 10.22, 17.75 ± 4.52 vs. 23.10 ± 7.13, respectively, all p values <.001). On the other hand left atrial reservoir longitudinal strain (LASr), conduit longitudinal strain (LAScd), contraction longitudinal strain (LASct), reservoir circumferential strain (LASr-c), conduit circumferential strain (LAScd-c), and contraction circumferential strain (LASct-c) were significantly lower in patient group (26.64 ± 5.64 vs.19.16 ± 4.58, -19.53 ± 3.72 vs. -11.28 ± 3.47, -10.34 ± 1.56 vs. -4.59 ± 1.49, 30.72 ± 4.04 vs. 19.31 ± 2.60, -19.91 ± 1.78 vs. -13.38 ± 2.85, -15.89 ± 6.37 vs. -9.24 ± 1.63, respectively, all p values <.001). The present study found that premature ventricular complexes can lead to atrial remodeling as well as ventricular remodeling in patients with PVC and 4D LAQ technology can quantitatively examine left atrial function and determine these alterations early.

Could premature ventricular contractions lead to atrial remodeling?

Abstract Background Premature ventricular contraction (PVC) is a frequent kind of arrhythmia that affects around 1% of the general population. While PVC most frequently impairs ventricular function in structurally normal hearts, retrograde ventriculo-atrial conduction can occur in people with PVC. These retrograde atrial activations may mimic pulmonary vein-derived atrial ectopies. As a result, PVC may raise the risk of AF by retrograde ventriculo-atrial conduction. The Four-Dimensional Automated Left Atrial Quantification (4D Auto LAQ) tool is a left atrial analytical approach that utilizes three-dimensional volume data to quantify the volume, as well as LA longitudinal and circumferential strain. The purpose of this study was to determine if clinical diagnosis of PVC is connected with abnormal LA functions as determined by LA strain evaluation utilizing a 4D Auto LAQ compared to the normal population. Methods 58 patients with frequent PVCs were enrolled in the research, along with 53 healthy volunteers as a control group. Imaging was performed using the GEvivid E95 ultrasonic diagnostic instrument (GE Healthcare; Vingmed Ultrasound, Horten, Norway) equipped with an M5S probe (frequency range: 1.5–4.6 MHz) and a 4V probe (frequency range: 1.5–4.0 MHz). Images were imported into and selected for analysis using the EchoPAC203 software (GE Healthcare). The analysis mode was selected, followed by the volume and 4D Auto LAQ submodes. Following that, the sample point was positioned in the center of the mitral orifice in each of the three planes. The'review' function was used to acquire the LA parameters measured by 4D Auto LAQ, including volume and strain parameters. Results The maximum left atrial volume (LAVmax) and minimal left atrial volume (LAVmin) were signifcantly higher in patient group (38.91±9.72 vs. 46.31±10.22, 17.75±4.52 vs. 23.10±7.13 respectively, all p values &amp;lt;0.001).On the other hand left atrial reservoir longitudinal strain (LASr), conduit longitudinal strain (LAScd), contraction longitudinal strain (LASct), reservoir circumferential strain (LASr-c), conduit circumferential strain (LAScd-c), and contraction circumferential strain (LASct-c) were signifcantly lower in patient group (26.64±5.64 vs. 19.16±4.58, −19.53±3.72 vs. −11.28±3.47, −10.34±1.56 vs. −4.59±1.49, 30.72±4.04 vs. 19.31±2.60, −19.91±1.78 vs. −13.38±2.85, −15.89±6.37 vs. −9.24±1.63, respectively all p values &amp;lt;0.001). Conclusion The present study found that premature ventricular contractions can lead to atrial remodeling as well as ventricular remodeling in patients with PVC and 4D LAQ technology can quantitatively evaluate atrial myocardial function and detect these changes early. Funding Acknowledgement Type of funding sources: None.

Improvement of Student Weariness Emotion in English Classroom Based on Intelligent Internet of Things and Big Data Technology.

In order to improve the recognition effect of student weariness emotion in English classroom, this paper combines intelligent Internet of Things technology and big data technology to construct an improvement model of student weariness emotion in English classroom. In the process of student facial expression recognition, according to the given grayscale threshold, this paper extracts the surface contour information from the three-dimensional volume data, extracts the student's surface contour information, and uses triangular facets to fit to form a triangular mesh. Moreover, this paper renders a triangular mesh model and shows how to speed up the calculation of PFH. In addition, this paper proposes a Fast Point Feature Histogram, which uses an iterative closest point fine registration algorithm for image registration. Finally, this paper constructs an emotion recognition model of students' weariness in English classroom. From the test results, it can be seen that the student weariness emotion recognition system in English classroom proposed in this paper can effectively identify students' weariness emotion.

A Handheld Fiber-Optic Probe to Enable Optical Coherence Tomography of Oral Soft Tissue.

This study presents a highly miniaturized, handheld probe developed for rapid assessment of soft tissue using optical coherencetomography (OCT). OCT is a non-invasive optical technology capable of visualizing the sub-surface structural changes that occur in soft tissue disease such as oral lichen planus. However, usage of OCT in the oral cavity has been limited, as the requirements for high-quality optical scanning have often resulted in probes that are heavy, unwieldy and clinically impractical. In this paper, we present a novel probe that combines an all-fiber optical design with a light-weight magnetic scanning mechanism to provide easy access to the oral cavity. The resulting probe is approximately the size of a pen (10 mm × 140 mm) and weighs only 10 grams. To demonstrate the feasibility and high image quality achieved with the probe, imaging was performed on the buccal mucosa and alveolar mucosa during routine clinical assessment of six patients diagnosed with oral lichen planus. Results show the loss of normal tissue structure within the lesion, and contrast this with the clear delineation of tissue layers in adjacent inconspicuous regions. The results also demonstrate the ability of the probe to acquire a three-dimensional data volume by manually sweeping across the surface of the mucosa. The findings of this study show the feasibility of using a small, lightweight probe to identify pathological features in oral soft tissue.

Automatic plane adjustment of orthopedic intraoperative flat panel detector CT-volumes.

Purpose: To assess the result in orthopedic trauma surgery, usually three-dimensional volume data of the treated region is acquired. With mobile C-arm systems, these acquisitions can be performed intraoperatively, reducing the number of required revision surgeries. However, the acquired volumes are typically not aligned to the anatomical regions. Thus, the multiplanar reconstructed (MPR) planes need to be adjusted manually during the review of the volume. To speed up and ease the workflow, an automatic parameterization of these planes is needed. Approach: We present a detailed study of multitask learning (MTL) regression networks to estimate the parameters of the MPR planes. First, various mathematical descriptions for rotation, including Euler angle, quaternion, and matrix representation, are revised. Then, two different MTL network architectures based on the PoseNet are compared with a single task learning network. Results: Using a matrix description rather than the Euler angle description, the accuracy of the regressed normals improves from 7.7deg to 7.3deg in the mean value for single anatomies. The multihead approach improves the regression of the plane position from 7.4 to 6.1mm, whereas the orientation does not benefit from this approach. Thus, the achieved accuracy meets the reported interrater variance in similarly complex body regions of up to 6.3deg for the normals and up to 9.3mm for the plane position. Conclusions: The use of a multihead approach with shared features leads to more accurate plane regression compared with the use of individual networks for each task. It also improves the angle estimation for the ankle region. The reported results are in the same range as manual plane adjustments. The use of a combined network with shared parameters requires less memory, which is a great benefit for the implementation of an application for the surgical environment.

HFCF-Net: A hybrid-feature cross fusion network for COVID-19 lesion segmentation from CT volumetric images.

BackgroundThe coronavirus disease 2019 (COVID‐19) spreads rapidly across the globe, seriously threatening the health of people all over the world. To reduce the diagnostic pressure of front‐line doctors, an accurate and automatic lesion segmentation method is highly desirable in clinic practice.PurposeMany proposed two‐dimensional (2D) methods for sliced‐based lesion segmentation cannot take full advantage of spatial information in the three‐dimensional (3D) volume data, resulting in limited segmentation performance. Three‐dimensional methods can utilize the spatial information but suffer from long training time and slow convergence speed. To solve these problems, we propose an end‐to‐end hybrid‐feature cross fusion network (HFCF‐Net) to fuse the 2D and 3D features at three scales for the accurate segmentation of COVID‐19 lesions.MethodsThe proposed HFCF‐Net incorporates 2D and 3D subnets to extract features within and between slices effectively. Then the cross fusion module is designed to bridge 2D and 3D decoders at the same scale to fuse both types of features. The module consists of three cross fusion blocks, each of which contains a prior fusion path and a context fusion path to jointly learn better lesion representations. The former aims to explicitly provide the 3D subnet with lesion‐related prior knowledge, and the latter utilizes the 3D context information as the attention guidance of the 2D subnet, which promotes the precise segmentation of the lesion regions. Furthermore, we explore an imbalance‐robust adaptive learning loss function that includes image‐level loss and pixel‐level loss to tackle the problems caused by the apparent imbalance between the proportions of the lesion and non‐lesion voxels, providing a learning strategy to dynamically adjust the learning focus between 2D and 3D branches during the training process for effective supervision.ResultExtensive experiments conducted on a publicly available dataset demonstrate that the proposed segmentation network significantly outperforms some state‐of‐the‐art methods for the COVID‐19 lesion segmentation, yielding a Dice similarity coefficient of 74.85%. The visual comparison of segmentation performance also proves the superiority of the proposed network in segmenting different‐sized lesions.ConclusionsIn this paper, we propose a novel HFCF‐Net for rapid and accurate COVID‐19 lesion segmentation from chest computed tomography volume data. It innovatively fuses hybrid features in a cross manner for lesion segmentation, aiming to utilize the advantages of 2D and 3D subnets to complement each other for enhancing the segmentation performance. Benefitting from the cross fusion mechanism, the proposed HFCF‐Net can segment the lesions more accurately with the knowledge acquired from both subnets.

A Study of Longitudinal NPS Measurement in CT Images Based on the Central Cross-section Theorem

Various approaches in noise power spectrum (NPS) analysis are currently used for measuring a patient's longitudinal (z-direction) NPS from three-dimensional (3D) CT volume data. The purpose of this study was to clarify the relationship between those NPSs and 3D-NPS based on the central slice theorem. We defined the 3D-NPS(fx, fy, fz) that was calculated by 3D Fourier transform (FT) from 3D noise data (3D-Noise(x, y, z), x-y scan plane). Here, fx, fy and fz are spatial frequencies corresponding to the axes of x, y and z, respectively. Based on the central slice theorem, we described three relationships as follows. (1) The fz-directional NPS calculated from the 3D-Noise(x=0, y=0, z) is equal to the profile obtained by projecting 3D-NPS(fx, fy, fz) in fx- and fy-directions. (2) The fz-directional NPS calculated from the profile obtained by projecting 3D-Noise(x=0, y, z) in the y-direction is equal to the profile at fy=0 in the data obtained by projecting 3D-NPS(fx, fy, fz) in the fx-direction. (3) The fz-directional NPS calculated from the profile obtained by projecting 3D-Noise(x, y, z) in x and y-directions is equal to the profile of 3D-NPS(fx=0, fy=0, fz). To verify them, we compared the NPSs measured from actual 3D noise data that were obtained using a cylindrical water phantom. In each relationship (1)-(3), the fz-directional NPS matched the profile obtained from the 3D-NPS(fx, fy, fz). Based on the central slice theorem, we clarified the relationships between fz-directional NPSs and 3D-NPS. We should understand them and then consider which method should be used for fz-directional NPS measurement.

A Study of 3D-NPS Analysis in CT Images Based on the Central Cross-section Theorem

The noise power spectrum (NPS) of a CT scanner is commonly measured from a single noise image. However, since CT images are three-dimensional (3D) volume data, they have 3D noise characteristics (3D-NPS). In this study, we clarify the relationship among NPSs measured by various approaches in NPS analysis based on the central slice theorem. Its validity is verified by the NPS measurements using actual 3D noise data. We defined the NPSz-projection(fx, fy) that was calculated by the 2D Fourier transform (FT) from the 2D projection of 3D noise data in the patient longitudinal direction, the 3D-NPS(fx, fy, fz) that was calculated by the 3D-FT from the 3D noise data, and the 2D-NPS(fx, fy) that was calculated by the 2D-FT from a single noise image; fx, fy, and fz are spatial frequencies corresponding to the axes of x, y, and z in the reconstructed CT volume, respectively. Based on the central slice theorem, we described that the NPSz-projection(fx, fy=0) was equal to the 3D-NPS(fx, fy=0, fz=0), and the NPS(2D-NPS(fx, fy=0)) was different from the 3D-NPS(fx, fy=0, fz=0). To verify them, we compared the NPSs calculated from actual 3D noise data that were obtained using a cylindrical water phantom. The 3D-NPS(fx, fy=0, fz=0) matched the NPSz-projection(fx, fy=0) and was different from the 2D-NPS(fx, fy=0). Based on the central slice theorem, we clarified the relationship among NPSs measured by various approaches in NPS analysis; it is important to understand this and then select an appropriate noise data handling and NPS measurement method.

Intelligent classification for three-dimensional metal powder particles

The shape of constituent metal particles has a significant influence on the bulk properties of raw powder and the mechanical properties of manufactured parts. In this paper, we provide a method for automatic shape-based classification of metal powder particles. Firstly, X-ray computed tomography was used to obtain three-dimensional volume data of metal powder particles, and a segmentation operation was performed to separate particles from one another. Classifying by particle shape, particles were manually labelled to create training and testing data, separating the powder particles into one of six user-defined categories: ‘connected’, ‘ellipsoidal’, ‘irregular’, ‘pear’, ‘porous’ and ‘spherical’. The machine learning network PointNet++ was used to classify particles, using 1024 automatically-defined geometric particle features, as well as twelve user-defined features. Our results show that the accuracy of this automatic classification method reaches 93.8%. Finally, an additional test batch of measured powder was used to verify our classification method.

Whole Three-Dimensional Dosimetry of Carbon Ion Beams with an MRI-Based Nanocomposite Fricke Gel Dosimeter Using Rapid T1 Mapping Method.

MRI-based gel dosimeters are attractive systems for the evaluation of complex dose distributions in radiotherapy. In particular, the nanocomposite Fricke gel dosimeter is one among a few dosimeters capable of accurately evaluating the dose distribution of heavy ion beams. In contrast, reduction of the scanning time is a challenging issue for the acquisition of three-dimensional volume data. In this study, we investigated a three-dimensional dose distribution measurement method for heavy ion beams using variable flip angle (VFA), which is expected to significantly reduce the MRI scanning time. Our findings clarified that the whole three-dimensional dose distribution could be evaluated within the conventional imaging time (20 min) and quality of one cross-section.

Craniofacial Reconstruction via Face Elevation Map Estimation Based on the Deep Convolution Neutral Network

In this study, to achieve the possibility of predicting face by skull automatically, we propose a craniofacial reconstruction method based on the end-to-end deep convolutional neural network. Three-dimensional volume data are obtained from 1447 head CT scans of Chinese people of different ages. The facial and skull surface data are projected onto two-dimensional space to generate a two-dimensional elevation map, and then, use the deep convolution neural network to realize the prediction of skull to face shape in two-dimensional space. The encoder and decoder are composed of first feature extraction through the encoder and then as the input of the decoder to generate the craniofacial restoration image. In order to accurately describe the features of different scales, we adopt an U-shaped codec structure with cross-layer connections. Therefore, the output features are decomposed with the features of the corresponding scales in the encoding stage to achieve the integration of different scales while restoring the feature scales in the compression and decoding stage. Meanwhile, the U-net structures help to avoid the problem of loss of detail features in the downsampling process. We use supervised learning to obtain the prediction model from skull to facial elevation map. Back-projection operation is performed afterwards to generate facial surface data in 3D space. Experiments show that the proposed method in this study can effectively achieve craniofacial reconstruction, and for most part of the face, restoration error is controlled within 2 mm.

Reliability of Sonography-based Volume Computer Aided Diagnosis in the Normal Fetal Heart.

To explore the inter- and intra-observer reliability of Sonography-based Volume Computer Aided Diagnosis (SonoVCAD) in the display of 8 diagnostic planes of fetal echocardiography and to evaluate its efficiency. Three-dimensional volume data sets of the 56 normal singleton fetuses were acquired from a 4-chamber view by using a volume probe. After processing the data sets by using SonoVCAD, 8 cardiac diagnostic planes were displayed automatically. Three doctors with different experiences of performing fetal echocardiography evaluated each diagnostic plane and the success rates of 8 diagnostic planes were calculated. Inter-observer and intra-observer reliabilities were estimated by Cohen's kappa statistics. A total of 276 volume data sets acquired from the 56 normal fetuses were used for SonoVCAD analysis and display. The success rate of each diagnostic section was more than 90%, ranging from 90.6% to 99.6%. Among 276 volumes, 81.5% (225/276) of volumes were able to generate all 8 diagnostic views successfully. Moderate to substantial agreement (kappa, 0.509-0.794) was found between 2 less experienced operators. Moderate to near-perfect agreement (kappa, 0.439-0.933) was found between an expert and 2 less experienced sonographers. Intra-observer reliability was substantial to near-perfect (kappa, 0.602-0.903). The efficiency of SonoVCAD was assessed. The expert spent less time than 2 less experienced examiners (P < 0.001) but no significant difference was found between 2 less experienced examiners (P = 0.176). Besides, SonoVCAD consumed significantly less time than 2-dimensional ultrasound (P < 0.001). SonoVCAD can significantly improve the success rates of 8 diagnostic planes in fetal echocardiography with low operator dependency, good reproducibility and high efficiency.

Research on Mechanical Deviation Correction Method for Industrial CT

Industrial CT is the abbreviation of industrial computed tomography (CT) technology, which adopts X-ray imaging principle to realize the non-destructive three-dimensional imaging to obtain high-precision three-dimensional tomographic data and volume data inside the product, clear, accurate and intuitive. In addition, it can show the internal structure, composition, material and defective condition of the inspected object. Factors such as mechanical accuracy and mounting deviations can adversely affect CT imaging and, in severe cases, lead to CT imaging failure. It is important to achieve effective correction based on the existing CT mechanical accuracy and mounting deviations. In this paper the authors collect data and performed experiments in a real CT system, they use standard test blocks to correct the deviations while keeping the existing CT system unchanged, the effectiveness can be validated by the experimental data.