Plane Images Research Articles

Tunable image quality control of 3-D ultrasound using switchable CycleGAN.

In contrast to 2-D ultrasound (US) for uniaxial plane imaging, a 3-D US imaging system can visualize a volume along three axial planes. This allows for a full view of the anatomy, which is useful for gynecological (GYN) and obstetrical (OB) applications. Unfortunately, the 3-D US has an inherent limitation in resolution compared to the 2-D US. In the case of 3-D US with a 3-D mechanical probe, for example, the image quality is comparable along the beam direction, but significant deterioration in image quality is often observed in the other two axial image planes. To address this, here we propose a novel unsupervised deep learning approach to improve 3-D US image quality. In particular, using unmatched high-quality 2-D US images as a reference, we trained a recently proposed switchable CycleGAN architecture so that every mapping plane in 3-D US can learn the image quality of 2-D US images. Thanks to the switchable architecture, our network can also provide real-time control of image enhancement level based on user preference, which is ideal for a user-centric scanner setup. Extensive experiments with clinical evaluation confirm that our method offers significantly improved image quality as well user-friendly flexibility.

Light Field Image Super-Resolution Network via Joint Spatial-Angular and Epipolar Information

This paper proposes a novel efficient convolutional neural network (CNN) for light field (LF) spatial super-resolution (SR). Due to the high dimensionality of LF data, it is important to decompose the original 4D LF data structure into a lower data subspace helping to reduce memory and computational costs. In this context, we propose a novel hybrid CNN structure working on different 2D data representations to fully handle the 4D structure. Our proposed network can simultaneously exploit information from three different types of 4D LF representations: spatial, angular, and epipolar plane image (EPI) information. Based on pixel similarities of each data subspace, we propose an inter-intra spatial feature extractor (II-SFE) and an inter-intra angular feature extractor (II-AFE) that can exploit the pixel correlations among the spatial-angular domain. Moreover, based on the correlated epipolar geometry among pixels, we also propose a multi-orientation epipolar feature extractor (MO-EFE) by decomposing an LF image into four orientation stacks of views and feeding them into different network branches to learn rich sub-pixel information in the EPI domain. The extracted features of each 2D subspace are integrated by our proposed attention fusion module (AFM), which adaptively rescales the embedding weights based on their correlations. Experimental results on both small- and wide-baseline benchmark LF datasets confirm the effectiveness of the proposed method compared to the well-known state-of-the-art methods in terms of objective and subjective performance.

Retracted: Optimization of Plane Image Color Enhancement Processing Based on Computer Vision Virtual Reality

Retracted: Optimization of Plane Image Color Enhancement Processing Based on Computer Vision Virtual Reality

A colour enhancement method of plane image based on the interactive interface of mobile smart devices

A colour enhancement method of plane image based on the interactive interface of mobile smart devices

A colour enhancement method of plane image based on the interactive interface of mobile smart devices

A colour enhancement method of plane image based on the interactive interface of mobile smart devices

Retracted: Optimization of Color Enhancement Processing for Plane Images Based on Computer Vision

Retracted: Optimization of Color Enhancement Processing for Plane Images Based on Computer Vision

Light-field angular super-resolution reconstruction via fusing 3D epipolar plane images

Light-field angular super-resolution reconstruction via fusing 3D epipolar plane images

Evaluation of Anatomical Variation of Human Coccyx on Magnetic Resonance Imaging

Introduction: The coccyx, a triangular bone forming the last part of the vertebral column, acts as the weight-bearing structure when a person is seated, and is the site of insertion of the pelvic floor tendons and numerous ligaments attached to it. The main objective of this study was to investigate the morphology and morphometry of the coccyx on lumbar spine MRI images in asymptomatic individuals among Nepalese adults. Methods: This study was conducted retrospectively on the lumbar spine images of 190 adult population without a history of trauma in the coccyx region, from April to September 2019. The coccygeal vertebrae count, the number of bone segments, and intercoccygeal and joint fusions were determined from the sagittal plane images. In addition, the length and angles were also measured. Results: One hundred and ninety patients, with a mean age of 43.91 years, were enrolled in the study. Among the 4 types of coccyx; the most common type was type II (46.3%), followed by type I (40.5%). The coccyx was formed from 4 vertebral segments in 21.1% (n=41) individuals, 3 vertebral segments in 45.3% (n=86) individuals, 2 vertebral segments in 31.1% (n=59) individuals and 1 vertebral segment in 2.1% (n=4) individuals. Conclusions: In our study, type II coccyxes predominated. The prevalence of sacrococcygeal and intercoccygeal fusion, as well as the number of coccygeal vertebrae, were similar in the Nepalese population and the Western population.

TMSO-Net: Texture adaptive multi-scale observation for light field image depth estimation

Light field can record the four-dimensional information of light rays, i.e. the position and direction information in which depth information is implied. To improve the depth estimation accuracy, we propose a depth estimation algorithm based on convolutional neural network (CNN). First, a single image super resolution algorithm is adopted to spatially super resolve the sub-aperture images (SAIs). Second, to adapt the texture complexity, the SAIs are partitioned into two regions, i.e., simple texture region and complex texture region, based on the texture analysis of the central SAI. Third, the epipolar plane images (EPIs) in horizontal, vertical, 45 degree diagonal, and 135 degree diagonal directions for both complex and simple texture regions are extracted, and the corresponding EPIs for the simple and complex texture regions are fed into the specified network branches. Finally, a fusion module is designed to generate the depth map. Experimental results show that the quality of the estimated depth maps by the proposed method is better than the state-of-the-art methods in terms of both objective quality and subjective quality. Moreover, the proposed method is more robust to noise.

Foramen tympanicum prevalence in the population of Southeast Brazil: a morphological study in computed tomography scans.

The foramen tympanicum is located on the anteroinferior region of the external acoustic meatus and posteromedial to the temporomandibular joint in children between the 1st and the 5th year of life. It is considered an anatomical variation when it persists in adults. The aim of this study was to verify the prevalence as well as to characterise the foramen tympanicum in computed tomography (CT) scans of the population from southeastern part of Brazil. A total of 78 CT scans of dry human skulls (20 female and 58 male) were used, which were selected randomly regarding the ages, ranged from 15 to 100 years. The foramen tympanicum was identified in the images of the axial plane and confirmed in the images of the coronal and sagittal planes. The largest diameter (in mm) was obtained. The descriptive statistics (in %), Fisher's test and χ2 test were performed to compare the prevalence of foramen tympanicum between sexes and sides. The probability value ≤ 0.05 was defined as a level of significance. Descriptive statistics were performed to verify the mean diameter of the foramen on the right and left sides of the skulls. The prevalence of foramen tympanicum was higher in females (p = 0.0070), bilaterally, as the absolute values of females were lower in relation to males. Fisher's exact test showed that the prevalence of foramen tympanicum was significantly higher in females (45%) than in males (15.52%). On the right side, the mean axial diameter was 2.23 mm (range 0.93-3.75 mm). On the left side, the mean axial diameter was 2.22 mm (range 0.9-3.61 mm). The knowledge of anatomical variations is extremely valuable for an accurate diagnosis, treatment plan and prognosis and a thorough preoperative assessment.

Visual Metaphor and Image Interpretation of “Guan-Yin-Ma-Lian”

From the perspective of production technology, the god image usually has two manifestations: paintings and sculptures, while there are two main forms of paintings: murals and prints. In this paper, the engravings of gods named “Guan-yin-ma-lian” (觀音媽聯) were taken as the research object. The origin of “Guan-yin-ma-lian” is traced back. The form of plane composition of this type of engraving is analyzed from the perspective of iconology. The image of Mazu (媽祖) and its implied meaning in “Guan-yin-ma-lian” are discussed from the perspective of folk beliefs. As one of the plane images of Mazu, “Guan-yin-ma-lian” is not only a very important religious art in folk culture, but also an important link closely connecting Mazu belief circles at home and abroad. It is also an important cultural asset of Chinese people who worship Mazu.

Height aware understanding of remote sensing images based on cross-task interaction

3D scene perception that provides semantic interpretation for each point has a wide range of practical scenarios. However, existing methods are generally implemented based on 3D point cloud. In this work, we investigate the inverse projection problem from 2D plane image to 3D scene perception and propose a multi-task model to jointly predict height information and semantic categories. Taking into account the commonalities and differences between the two tasks, we have carefully designed the multi-task model to explicitly strengthen the establishment of cross-task correlations. Specifically, the calibration refinement attention (CRA) module is proposed before the classifier heads, incorporating the beneficial information while filtering the inconsistent characteristics among the two tasks. Besides, a spatial structure enhanced (SSE) module is introduced to integrate the spatial structure information into the output features of CRA module through skip-connection. After that, the neighboring pixel affinity (NPA) loss and the soft weighted ordinal (SWO) classification loss for the two tasks are introduced to optimize the direction of the task gradients. At the same time, to validate the effectiveness of the proposed method, a novel metric named height constrained semantic accuracy (HCSA) is proposed to consider the accuracy of semantic segmentation and height estimation jointly. Extensive experiments on Vaihingen, Potsdam, and DFC2019 demonstrate that the proposed method achieves generalization and competitive performance.

Optimized auto-focusing method for 3D ultrasound imaging in NDT

In this work, we propose an optimized auto-focusing algorithm to be used with ultrasound matrix arrays for three-dimensional imaging in the presence of interfaces. Its main application is for non-destructive-testing (NDT), where usually two mediums with different propagation velocities are present. The refraction at the interface between those mediums complicates the calculation and the application of the focusing delays, which requires iterative processes that are usually the main bottleneck for high-speed inspections.We tackle this problem by generalizing the virtual array concept previously proposed for linear arrays and plane images to the case of matrix-arrays and three-dimensional (3D) beamforming. Furthermore, we propose a two-dimensional iterative algorithm for exact time-of-flight calculation to the two foci per scan line required by the virtual array. The mathematical formulation for the 3D case is obtained and the time-delay errors are analyzed by simulation. Finally, image quality with the proposed method is compared with that obtained with conventional Total Focusing Method (TFM) and exact delay calculation. The principal conclusion is that the virtual array approach generates images with equivalent quality to those obtained by exact calculation, while it reduces more than two orders of magnitude the computation load.

Effect of anterior maxillary wall shape on ease of prelacrimal recess approach among Asians and Westerners.

Endoscopic sinus surgery has been widely used in the treatment of most maxillary diseases, although the inability to access lesions in the anterior and inferior maxillary sinus walls is a major disadvantage. In few cases, the prelacrimal recess (PLR) approach can be used, which secures an overall view within the maxillary sinus. The distance between the anterior maxillary wall and lacrimal duct is related to the ease of the PLR approach. First we measured the distance between the anterior maxillary wall and lacrimal duct using anatomical image analysis to classify the types and then evaluated whether anatomical factors were related to the results. Both sides of the sinuses were evaluated in 272 participants (544 sides). After marking a tangent line (line 1 [L1]) through the posterior surface of the anterior maxillary sinus wall and a parallel line (line 2 [L2]) to the anterior surface of the lacrimal duct, the vertical distance between L1 and L2 was measured. Vertical distances of < 3mm, 3-7mm, and > 7mm were classified as PLR approach types I, II, and III, respectively. In the axial plane image, line 3 (L3) (a horizontal line starting from the inner anterior maxillary sinus wall) was drawn and the angle with L1 (L1-L3A) was measured. The proportions of types I, II, and III were 23.2% (126), 55.0% (299), and 21.8% (119), respectively. The mean L1-L3As for types I, II, and III were 12.87 ± 4.92°, 11.20 ± 5.08°, and 10.40 ± 4.47°, respectively, showing a significant difference in mean values (p < 0.001). The L1-L3A and vertical distance between L1 and L2 showed a significant negative correlation (r = - 0.201, p < 0.001). We observed a correlation between the distance from the anterior maxillary wall to the lacrimal duct and L1-L3A. The L1-L3A indicates the degree of curvature of the anterior maxillary wall; therefore, the smaller the L1-L3A, the easier it may be to access the PLR.

Identification and evaluation of spalling aggregate in chip seals in three dimensions

In the context of the present lack of aggregate spalling identification methods and the imprecise recognition of images of aggregate spalling points owing to aggregate stacking and coverage in the course of chip-seal paving, this paper proposes an aggregate spalling recognition method based on 3-D laser elevation data points in order to improve the recognition accuracy of spalling aggregate in chip seals. In this study, 3-D point elevation data on the surface of a chip seal specimen were fed into the 3-D laser detection equipment, the raw data were smoothed, the super four-point fast and robust matching method was used on the 3-D laser points for registration, and 3-D reconstructions of the chip seal were built based on the laser point after registration, and 3-D reconstructions before and after spalling were overlaid with the different values of the surface elevations before and after spalling. The spalling plane curve was then obtained using contours from the differences in the values of the surface elevation before and after spalling, and the spalling perimeter Ln, spalling area Sn, and spalling volume Vn were then calculated on the basis of the 3-D coordinates of the spalling curve. The results showed that the relative errors of the spalling area measured by the laser method and the OTSU method (an image segmentation algorithm proposed by Otsu (1979)) were #1 3.1% vs 4.0%, #2 4.4% vs 6.4%, #3 3.3% vs 58.2%, respectively, indicating that the bitumen attached to the aggregate and the overlap of plane images of aggregate due to aggregate stacking are the main reasons for the greater errors in the recognition of image method. The relative errors in the spalling perimeter of the three specimens are 12.61%, 1.60%, and 0.94%, the relative errors of spalling area are 3.13%, 4.36%, and 3.31%, and the relative errors of spalling volume are 6.69%, 6.57%, and 8.71%, which are less than the approximately same level 10% indicating that the accuracy and the stability of aggregate spalling recognition based on 3-D laser elevation data is excellent.

Classification and Object Detection of 360° Omnidirectional Images Based on Continuity-Distortion Processing and Attention Mechanism

The use of 360° omnidirectional images has occurred widely in areas where comprehensive visual information is required due to their large visual field coverage. However, many extant convolutional neural networks based on 360° omnidirectional images have not performed well in computer vision tasks. This occurs because 360° omnidirectional images are processed into plane images by equirectangular projection, which generates discontinuities at the edges and can result in serious distortion. At present, most methods to alleviate these problems are based on multi-projection and resampling, which can result in huge computational overhead. Therefore, a novel edge continuity distortion-aware block (ECDAB) for 360° omnidirectional images is proposed here, which prevents the discontinuity of edges and distortion by recombining and segmenting features. To further improve the performance of the network, a novel convolutional row-column attention block (CRCAB) is also proposed. CRCAB captures row-to-row and column-to-column dependencies to aggregate global information, enabling stronger representation of the extracted features. Moreover, to reduce the memory overhead of CRCAB, we propose an improved convolutional row-column attention block (ICRCAB), which can adjust the number of vectors in the row-column direction. Finally, to verify the effectiveness of the proposed networks, we conducted experiments on both traditional images and 360° omnidirectional image datasets. The experimental results demonstrated that better performance than for the baseline model was obtained by the network using ECDAB or CRCAB.

Low-Divergence hBN Single-Photon Source with a 3D-Printed Low-Fluorescence Elliptical Polymer Microlens.

Efficiently collecting light from single-photon emitters is crucial for photonic quantum technologies. Here, we develop and use an ultralow fluorescence photopolymer to three-dimensionally print micrometer-sized elliptical lenses on individual precharacterized single-photon emitters in hexagonal boron nitride (hBN) nanocrystals, operating in the visible regime. The elliptical lens design beams the light highly efficiently into the far field, rendering bulky objective lenses obsolete. Using back focal plane imaging, we confirm that the emission is collimated to a narrow low-divergence beam with a half width at half-maximum of 2.2°. Using photon correlation measurements, we demonstrate that the single-photon character remains undisturbed by the polymer lens. The strongly directed emission and increased collection efficiency is highly beneficial for quantum optical experiments. Furthermore, our approach paves the way for a highly parallel fiber-based detection of single photons from hBN nanocrystals.

Direct model-based wavefront sensorless method with a fixed number of measurements

In wavefront sensorless (WFSL) adaptive optics, the intensity image in the observation plane, instead of the wavefront sensor, is utilised to estimate the input aberration. The number of intensity measurements is critical for applications with ever-changing phase aberration, such as astronomical imaging. This paper details two direct WFSL methods that need a fixed number of intensity measurements to estimate the input aberration. The proposed methods adopt a zonal approach rather than a modal one to estimate the phase aberration. Simulation results demonstrate that after applying the proposed methods, the aberration correction percentage can rise by approximately 70% for large aberrations.

Theme and Variation: Structuring Thrombus Formation from Jugular/Arterial Puncture Wounds to Occlusive Clots in a Mouse Model

Hemostasis in veins and arteries has been assumed to be different processes, e.g., white vs. red clots. We took the unbiased experimental approaches of 3D EM across full clot volumes and high-resolution heat mapping of platelet activation state to probe for commonalities and variations using controlled mouse models. Remarkably, the thrombi formed, 5 min post-puncture, in the jugular vein and femoral artery showed considerable consistency in organization. There were vaulted cavities within the puncture wound and apparent extravascular capping of the puncture hole to produce bleeding cessation. Notable differences in the venous thrombi were a greater activation of platelets lining the surfaces of the vaulted columns, an outcome that we attribute to lowr flow rates resulting in the longer retention of activating signals, and in the "hanging drop” nature of the arterial thrombi anchored to the extrema with an outward, extravascular bulge of the thrombus. We attribute these variations to arterial vessel wall properties and higher flow rates. At 20 min post-puncture, the intravascular crown of the venous wound thrombi was enriched in loosely adherent platelets in a low activation state. Similarly, the outer portion of the intravascular arterial thrombus was rich in low activation platelets including obviously discoid platelets. Strikingly, the arterial thrombus was subject to rebleeding and ate 20 min displayed fractured extravascular features indicative of an unstable clot. The puncture wound thrombi form in the relatively small volume of the immediate wound site. In contrast, occlusive clots fill the cylindrical volume of the vessel. To us, the overall structure of the wound thrombi suggested a predictive hypothesis for induced occlusive clotting. We propose that 1) the exposed intima lining the arterial vessel wall provides a cylindrical substratum for the generation of a tightly adherent platelet layer anchoring the forming clot and 2) the central volume of the clot resembles the intravascular crown of the puncture wound thrombus. Hence, the prediction of a central volume of the occlusive clot consisting of loosely adherent to discoid shaped platelets. That was result found for FeCl3 induced femoral clots by wide area-TEM (WA-TEM) i.e., a central volume rich in discoid shaped platelet and few red blood cells were trapped in the tightly adherent rim of the clot.. Using sequential block face SEM, we have begun to image and volume render carotid clots by sequential block face SEM. Analysis of 5,000 sequential images from the downstream half of the carotid clot, showed that the structure gradually tapered in a conical manner and contains few red blood cells. We propose that there are indeed organizational themes shared by the puncture wound thrombus in venous or arterial settings and the occlusive clot in which vascular damage is restricted to the endothelial cell layer. Work was supported by NIH grants (BS) R01 HL155519 and P20 HL146373 (Joel Bennett, principal investigator), R35 HL150118 (SWW) and by intramural NIBIB research program (RDL). An occlusive FeCl3 clot in the carotid artery of a C57BL/6 mouse imaged by sequential block face SEM over a length of 1.2 millimeters.A. cross section of the clot 200 microns in from the mid-occlusion starting point for sectioning. B. Zoom of A) to illustrate the low level of trapped red blood cells within the clot. C. cross section of the clot 900 microns in from the starting point showing how the clot has now tapered down and fills a minor portion of the artery. D. Rendering of 1.2 millimeters of the clot (blue), vessel wall (green) and in gray a single plane image. Flow was from left to right. Note that the clot progressively tapers to a point in the downstream direction.Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Individual differences in processing ability to transform visual stimuli during the mental rotation task are closely related to individual motor adaptation ability.

Mental rotation (MR) is a well-established experimental paradigm for exploring human spatial ability. Although MR tasks are assumed to be involved in several cognitive processes, it remains unclear which cognitive processes are related to the individual ability of motor adaptation. Therefore, we aimed to elucidate the relationship between the response time (RT) of MR using body parts and the adaptive motor learning capability of gait. In the MR task, dorsal hand, palmar plane, dorsal foot, and plantar plane images rotated in 45° increments were utilized to measure the RTs required for judging hand/foot laterality. A split-belt treadmill paradigm was applied, and the number of strides until the value of the asymmetrical ground reaction force reached a steady state was calculated to evaluate the individual motor adaptation ability. No significant relationship was found between the mean RT of the egocentric perspectives (0°, 45°, and 315°) or allocentric perspectives (135°, 180°, and 225°) and adaptive learning ability of gait, irrespective of body parts or image planes. Contrarily, the change rate of RTs obtained by subtracting the RT of the egocentric perspective from that of the allocentric perspective in dorsal hand/foot images that reflect the time to mentally transform a rotated visual stimulus correlated only with adaptive learning ability. Interestingly, the change rate of RTs calculated using the palmar and plantar images, assumed to reflect the three-dimensional transformation process, was not correlated. These findings suggest that individual differences in the processing capability of visual stimuli during the transformation process involved in the pure motor simulation of MR tasks are precisely related to individual motor adaptation ability.