Search

High tibial osteotomy (HTO) is a knee preservation procedure used to treat osteoarthritis of the knee. Identifying the hinge point, surgical point, and Fujisawa point in the patient's knee X-ray before surgery is a critical task. The aim of this study was to propose an artificial intelligence-based method to effectively help surgeons select the location of these landmark points, which provides important reference for subsequent surgery.We proposed PFDNet (parallel fusion decoding network), a novel convolutional neural network for key-point estimation of knee X-rays. PFDNet employs Res2Net for feature extraction in the network encoding phase and two partial decoders connected in parallel in the network decoding phase to finely aggregate the multiscale feature information produced by Res2Net. A total of 1842 knee X-ray images were trained, validated and predicted by PFDNet to determine whether the network could accurately detect key-points in the HTO surgical plan.At the hinge point, surgical point, and Fujisawa point, the average error and standard deviation from the calibration value in the PFDNet test results were 2.06 ± 1.165 mm, 2.713 ± 1.457 mm, and 2.015 ± 1.304 mm, respectively. This method exhibits superior performance compared with four convolutional neural network models that are also based on encoding and decoding frameworks: U-Net, ResUnet, SegNet, and FCN.The hinge point, surgical point, and Fujisawa point can be clearly selected by PFDNet from knee X-ray images and is locked to the millimeter level. The results show that the proposed artificial intelligence-based strategy can be instrumental in preoperative HTO planning.

The polysaccharides were isolated from apple pomace by hot-water extraction, and their anti-fatigue activity was evaluated in C2C12 muscle myoblasts and male Kunming mice. The purified polysaccharides from apple pomace (PAP) have a molecular weight of 1.74 × 10

for particle image velocimetry (PIV) technique, the two-dimensional (2D) PIV by one camera can only obtain 2D velocity field, while three-dimensional (3D) PIV based on tomography by three or four cameras is always complex and expensive. In this work, a binocular-PIV technology based on two cameras was proposed to reconstruct the 3D velocity field of gas-liquid two-phase flow, which is a combination of the binocular stereo vision and cross-correlation based on fast Fourier transform (CC-FFT). The depth of particle was calculated by binocular stereo vision on space scale, and the plane displacement of particles was acquired by CC-FFT on time scale. Experimental results have proved the effectiveness of the proposed method in 3D reconstruction of velocity field for gas-liquid two-phase flow.

An early warning method of compressor valve fault based on multi-parameter signals (vibration, pressure, and temperature) is presented in this work. Due to the complexity working condition, the run data of the compressor are of problems like noise and feature aliasing, which makes it difficult to extract useful features and find out the running law from the original signals. In this work, an improved deep learning network Multi-Level Fusion long short-term memory based on Component Evaluating Empirical Mode Decomposition and Fuzzy C-Means (CEEMD-FCM & MLF-LSTM) for parameter prediction of reciprocating compressor and an information fusion strategy is proposed for compressor valve fault warning. The CEEMD-FCM & MLF-LSTM network consists of data processing block, information learning block, and prediction output block, which is mainly responsible for parameter prediction. In the data processing block, the CEEMD-FCM algorithm is used for parameter decomposition, noise removal, and fuzzy mode (FM) reconstruction, which generates the input for the information learning block to ensure the predicting accuracy and reduce model complexity. MLF-LSTM is constructed to predict the parameter in the future by learning the temporal and spatial characteristics of FMs of the run data. Then, an early warning strategy for compressor valve fault based on multi-source information fusion is developed. Experimental results have verified that the proposed CEEMD-FCM & MLF-LSTM model and early warning strategy could realize early warning of compressor valve fault effectively.

Recent approaches based on convolutional neural networks significantly improve the performance of structured light image depth estimation in fringe projection and speckle projection 3D measurement. However, it remains challenging to simultaneously preserve the global structure and local details of objects for the structured light images in complex scenes. In this paper, we design a parallel CNN-transformer network (PCTNet), which consists of a CNN branch, a transformer branch, a bidirectional feature fusion module (BFFM), and a cross-feature multi-scale fusion module (CFMS). The BFFM and CFMS modules are proposed to fuse local and global features of the double branches in order to achieve better depth estimation. Comprehensive experiments are conducted to evaluate our model on four structured light datasets, i.e. our established simulated fringe and speckle structured light datasets, and public real fringe and speckle structured light datasets. Experiments demonstrate that the proposed PCTNet is an effective architecture, achieving state-of-the-art performance in both qualitative and quantitative evaluation.

Aiming at the problem of the loss of 3D point cloud, due to the occlusion of the field of view in the 3D measurement process, a measurement scheme based on line-structured light with dual-camera is given. In addition, in the line-structured light measurement technology, the traditional light plane calibration is more complicated, and the three-dimensional measurement accuracy is relatively low. For this reason, this paper used the binocular polar line constraint to calibrate the physical parameters of the light plane. Experimental results show that the dual-camera measurement system can obtain high-precision global measurement results. The maximum measurement error is 0.091 51 mm, and the average measurement error is 0.076 05 mm. Compared with the traditional binocular matching method and the traditional laser triangulation method, this method can deal with the problem of field occlusion more effectively, thereby reducing the loss of measurement information in the measurement process.

The three-dimensional reconstruction of bubble flow field is of great significance to study the motion of gas-liquid two phase flow. Combined with the abundant information of bubbles in the light field images, a fast and accurate 3D reconstruction method for dilute bubbly flow based on the deep learning algorithm DIF-LeNet (Double Information Fusion with LeNet5 net) is proposed in this work. The calibration method and bubble segmentation algorithm of light field image for data processing is detailed. DIF-LeNet realizes the fusion of different dimension data and regression prediction. By DIF-LeNet, bubble depth could be predicted by fusing the refocused bubble image with the focal distance of the refocused image. Then, the three-dimensional bubble field could be reconstructed with the bubble depth, bubble center and bubble diameter. Comparing with the reconstruction method based on statistics and LeNet5, the reconstruction accuracy and speed of the proposed method based on DIF-LeNet are improved. Especially for the bubble without focused image in the refocused image sequence, the effect of 3D reconstruction based on DIF-LeNet is much prominent.

In the task of using robots to capture end-of-life cars, the position of the vehicle door frame needs to be grasped. A fast and accurate positioning of the vehicle door frame is the key to realizing the automatic car grasping process. Traditional methods for locating and grasping scrap cars rely heavily on manual operations and suffer from low grasping efficiency and poor accuracy. Therefore, this paper proposes a binocular vision robot vehicle door frame spatial localization method based on the improved YOLOv4. This method includes a lightweight and efficient feature fusion target detection network in complex environments, and the target detection results are combined with an enhanced SURF feature–matching method to locate the vehicle door frame position. To simplify the network structure, MobileNetv3 is used instead of the backbone network CSPDarknet53, and deep separable convolution is used in the network. To increase the sensitivity of the network to vehicle door frame targets in complex environments, an improved convolutional block attention module is added to the pyramid attention with simple network backbones. Moreover, adaptive spatial feature fusion is introduced into the network to fully use the features at different scales for more effective feature fusion. Compared with YOLOv4, the number of network parameters is reduced by 73.8%, the mAP is improved by 1.35%, and the detection speed is increased by 28.7%. The experimental results demonstrate that the positioning accuracy of the system is 0.745 mm, which meets the positioning measurement error of less than 1 cm required for the vehicle door frame. The paper also compares our findings with other network models. The results show that the method achieves a good balance between detection speed and detection accuracy, satisfying the task of identifying vehicle door frames in complex environments with good detection results.

Knowledge graph (KG) plays an increasingly important role to improve the recommendation performance and interpretability. A recent technical trend is to design end-to-end models based on information propagation schemes. However, existing propagation-based methods fail to (1) model the underlying hierarchical structures and relations, and (2) capture the high-order collaborative signals of items for learning high-quality user and item representations. In this paper, we propose a new model, called Hierarchy-Aware Knowledge Gated Network (HAKG), to tackle the aforementioned problems. Technically, we model users and items (that are captured by a user-item graph), as well as entities and relations (that are captured in a KG) in hyperbolic space, and design a hyperbolic aggregation scheme to gather relational contexts over KG. Meanwhile, we introduce a novel angle constraint to preserve characteristics of items in the embedding space. Furthermore, we propose a dual item embeddings design to represent and propagate collaborative signals and knowledge associations separately, and leverage the gated aggregation to distill discriminative information for better capturing user behavior patterns. Experimental results on three benchmark datasets show that, HAKG achieves significant improvement over the state-of-the-art methods like CKAN, Hyper-Know, and KGIN. Further analyses on the learned hyperbolic embeddings confirm that HAKG offers meaningful insights into the hierarchies of data., Accept to SIGIR2022

A knowledge graph (KG) consists of a set of interconnected typed entities and their attributes. Recently, KGs are popularly used as the auxiliary information to enable more accurate, explainable, and diverse user preference recommendations. Specifically, existing KG-based recommendation methods target modeling high-order relations/dependencies from long connectivity user-item interactions hidden in KG. However, most of them ignore the cold-start problems (i.e., user cold-start and item cold-start) of recommendation analytics, which restricts their performance in scenarios when involving new users or new items. Inspired by the success of meta-learning on scarce training samples, we propose a novel meta-learning based framework called MetaKG, which encompasses a collaborative-aware meta learner and a knowledge-aware meta learner, to capture meta users' preference and entities' knowledge for cold-start recommendations. The collaborative-aware meta learner aims to locally aggregate user preferences for each user preference learning task. In contrast, the knowledge-aware meta learner is to globally generalize knowledge representation across different user preference learning tasks. Guided by two meta learners, MetaKG can effectively capture the high-order collaborative relations and semantic representations, which could be easily adapted to cold-start scenarios. Besides, we devise a novel adaptive task scheduler which can adaptively select the informative tasks for meta learning in order to prevent the model from being corrupted by noisy tasks. Extensive experiments on various cold-start scenarios using three real data sets demonstrate that our presented MetaKG outperforms all the existing state-of-the-art competitors in terms of effectiveness, efficiency, and scalability., Accepted to TKDE2022

Phase encoding and phase-shift profilometry are two commonly used 3D measurement techniques. However, the acquired phases in the techniques are subject to jump errors due to phase ambiguity and phase errors caused by multiple heterodyne. The phase-shifting profilometry also makes the selection of fringe period difficult. To overcome this problem and achieve high-precision measurement, a phase unwrapping method that combines dual-frequency heterodyne with double complementary phase encoding is proposed. First, two wrapped phases are obtained by two groups of sinusoidal fringes; the heterodyne phase is obtained after heterodyne processing, and the high-frequency phase is expanded by heterodyne phase. Second, the fringe levels are obtained using the complementary phase encoding fringes that are shifted by half an order, and then the absolute phase is obtained by selecting different phase coding levels according to different regions for the first phase unwrapping; Finally, the phase noise is removed by exploiting the difference between the phase slopes of adjacent pixels. Experimental results show that a system with the proposed method achieves an RMS error of 0.015 mm. In addition, the period of dual-frequency heterodyne synthesis does not need to cover the whole field of view, which breaks the limitation of frequency selection of the traditional dual-frequency heterodyne method and triple frequency heterodyne method, enabling high-precision measurement with higher frequency fringes. This method overcomes the limitations of the phase principal value error when using higher frequency fringes for high-precision measurement, improves the measurement effect of reflective objects, and effectively avoids the error caused by phase jump.

The UHV eight-wire lifter is an important tool in frequent use for the construction and maintenance of UHV transmission lines, which is made of titanium alloy and weighs about 40 kg. Due to high-weight, the work efficiency is greatly reduced. In this paper, the design of a lightweight UHV eight-wire lifter is proposed by using carbon fiber cladding structure. The effects of three factors on the maximum load-bearing load of the carbon fiber exterior cladding structural member are excavated through the failed experiment of the carbon fiber exterior cladding tensile and shear test pieces and the finite element progressive damage numerical calculation. The orthogonal experiment is employed to obtain the optimal cladding structure of the eight-wire lifter with a carbon fiber exterior coating cladding. After that, carbon fiber eight-wire lifter is developed and subjected to mechanical tests. The test results show that the performance of carbon fiber eight-wire lifter meets the mechanical performance requirements of eight-wire lifter. Compared with the titanium alloy eight-wire lifter, the maximum load-bearing load of the carbon fiber eight-wire lifter is increased by 20.5%, and the weight is decreased by 37.5%.

Wheel measurement and positioning technology is an important technology in vehicle production, and is of great importance for the safety of vehicles. At present, visual measurement and other methods are widely used in automotive production and inspection, but these methods are limited to the measurement of regular-sized vehicles. There is no effective solution to the problem of wheel alignment for large special-purpose vehicles with multiple tires. In order to solve the wheel positioning problem of large-size special-purpose vehicles, this paper designs a vision measurement system for wheel parameters in large scenes by combining vision sensors with linear motion guides to extend the vision measurement range and complete the global calibration of multiple-vision sensors with the help of laser trackers and self-researched stereo targets. This paper proposes an Elementary–Advance global calibration algorithm to improve the accuracy of coordinate system conversion during global calibration. This paper establishes a correction model for the measurement errors that exist in the initial global calibration process, and iterates the objective function to optimize the coordinate system conversion parameters between multiple vision sensors to improve the measurement accuracy of the system. The experimental results show that the goodness of fit and the accuracy of fit for the feature cylinder are 98.31% and 99.03% based on the global calibration method of this paper. The standard deviation of measurements for the standard ruler of the Inva alloy is less than 0.391 mm in the large dimensional range of 6050 mm × 3500 mm × 800 mm. Finally, by comparing the measurement results with the TrukCam four-wheel alignment instrument, the feasibility of the visual measurement system designed in this paper for large scenarios of wheel parameters is verified, and the measurement accuracy requirements for four-wheel alignment of special-purpose vehicles are met.

Near Infrared spectroscopy (NIRS) qualitative analysis technology has shown excellent development potential in the field of blend fabrics. However, the qualitative detection method based on the convolutional neural network (CNN) is difficult to accurately extract the feature of the spectral data, which will lead to missing detection or false detection; when using deep learning to build a qualitative detection model, due to interference of the external environment and other factors, the spectral data collected may have outliers, this means that the knowledge generalization on anomalous testing data, which may have a different distribution of that of the training set, is not trivial, which will also lead to missing detection or false detection. To solve the above problems, this paper proposes a novel qualitative detection neural network by analyzing the near infrared spectral data of blend fabrics. Firstly, we remove the convolutional layer and pooling layer of the CNN, making full use of the feature to enhance the feature representation ability of the model. Secondly, adding the L1 norm of the feature coefficients as a penalty term to the loss function to force those features with high redundancy to become weaker. Thirdly, in order to improve the recognition accuracy of the anomalous spectral data and minimize the model uncertainty, an ensemble machine learning approach utilizing 5 neural networks in parallel is used. To show the superiority of our proposed method, the existing methods are used as competitive methods to compare with our method. Our homemade dataset contains 3482 samples of blend fabrics with 9 different compositions. The results show that the Micro-F1-score, Micro-Specificity, Weight-F1-score, and Weight-Specificity of this method respectively 99.71%, 99.96%, 99.73%, and 99.99%, the results further confirm the method has higher analysis accuracy and stability. In addition, the method proposed in this paper can greatly improve the recognition accuracy of the anomalous spectral data. It has important practical value in the qualitative detection of blend fabrics.

Multi-parameter regularization can improve the over-regularization or under-regularization in the regularized inversion of dynamic light scattering (DLS) data. However, the regularization parameter selected by a fixed adjustment factor is not independent of the other parameters, which results in all the selected parameters being affected by the singular value of the truncation point. This leads to oscillations and false peaks in the recovered particle size distribution (PSD) as the noise in the data increases. In this paper, the singular value distribution characteristics are investigated, and this leads to the proposal of a novel multi-parameter selection method. Using the proportional relationship between two adjacent singular values, a regular parameter function is constructed for the parameter selection. Parameter optimization is performed using fixed point iteration to obtain the regularization parameter sequence corresponding to the singular value distribution. Thus, the effects of small singular values on noise are suppressed. Simulated DLS data for monomodal, closely-spaced bimodal, widely-spaced bimodal and trimodal PSDs were inverted under a range of different noise levels. The results show that, the proposed method greatly reduces the oscillations and false peaks and significantly improves the resolution of the peak position in the recovered PSDs as the noise in the data increases. The performance of the method was also verified using experimental DLS data.

Intestinal fibrosis is an excessive proliferation of myofibroblasts and deposition of collagen, a condition frequently seen in Crohn's disease (CD). The mechanism underlying myofibroblast hyper-proliferation in CD needs to be better understood. In this report, we found that mTOR inhibitor rapamycin or mTOR deletion in CX3Cr1+ mononuclear phagocytes inhibits expression of interleukin (IL)-23, accompanied by reduced intestinal production of IL-22 and ameliorated fibrosis in the TNBS-induced fibrosis mouse model. This inhibition of IL-23 expression is associated with elevated autophagy activity. Ablating the autophagy gene Atg7 increases the expression of IL-23, leading to increased expression of IL-22 and increased fibrosis. Both induction of IL-22 and intestinal fibrosis occurred in RAG-/- mice and depletion of innate lymphoid cells (ILCs) attenuates the fibrotic reaction, suggesting that the pro-fibrotic process is independent of T and B cells. Moreover, IL-22 facilitates the transformation of fibroblasts into myofibroblasts. Finally, the fibrotic reaction was attenuated upon neutralization of either IL-23 or IL-22. Altogether, this study elucidated a signaling cascade underlying intestinal fibrosis in which altered mTOR/autophagy in CX3Cr1+ mononuclear phagocytes up-regulates the IL-23/IL-22 axis, leading to an excessive fibrotic response. Thus, our findings suggest that this cascade could be a therapeutic target for alleviation of CD fibrosis.

This paper presents an approach to implement multi-parameter (i.e., pressure, temperature, vibration, current, and liquid level) signals for fault diagnosis of the reciprocating compressor (RC). Due to the complexity of structure and motion of such compressor, the acquired signals involve transient impacts and noises. This causes the useful information to be corrupted and makes it difficult to diagnose the fault patterns accurately. A component estimating empirical mode decomposition (CEEMD) method is proposed to remove the random noise and improve data quality. Furthermore, a new template matching algorithm called de-dimension template with double-loop correction (DDT-DLC) is applied to diagnose the fault pattern contained in the time series signals. The DDT employs a judging criterion for key characterization parameters extraction and a multicellular parameter fusion method to reduce the dimension of the matching template, and then, the DLC supplies a double-loop correction algorithm to build a parameter state array computing model of the time series data by adjusting the dynamic factors. The proposed approach is validated with three fault patterns and the healthy pattern in a two-stage reciprocating air compressor. To confirm the superiority of the proposed method, its performance is compared with that of the traditional methods. The results have indicated that the proposed approach is of highly diagnostic accuracy and shortly computing time in the fault diagnosis.

Metal component surfaces are random textured and non-smooth. There are many stains on the surface of metal component that are similar to the gray scale of the scratches. The scratches have non-uniform gray distribution, various shapes, and low contrast in their background, posing challenges in accurate scratch detection. This paper presents a method for detecting weak scratches on metal component surfaces based on deep convolutional neural networks (DCNNs). First, a DCNN is trained using labeled scratch images. Then, the scratches and some faults are detected by the trained DCNN, and most of the faults can be removed through properly thresholding based on the size of connected regions. Finally, the scratch length united in the number of pixels is obtained by the skeleton extraction. The experimental results show that the proposed method can effectively deal with background noise, thereby achieving accurate scratch detection.

To quickly obtain accurate 3D data of dental cast model, this paper proposes a 3D reconstruction method for dental cast model based on structured light. This method combines the structured light with the motor turntable to obtain a group of 3D data for the dental cast model from multiple angles, and automatically registers the dental 3D data from multiple angles through the ball calibration of turntable. Compared with the real data of the dental cast model, the maximum error of the 3D reconstruction results in this paper is 0.115 mm. The reconstruction time of this process is about 130 s. The experimental results show that the method has high precision and high scanning speed for the 3D reconstruction of the dental cast model.

During intestinal inflammation, immature cells within the intestinal crypt are called upon to replenish lost epithelial cell populations, promote tissue regeneration, and restore barrier integrity. Inflammatory mediators including TH1/TH17-associated cytokines influence tissue health and regenerative processes, yet how these cytokines directly influence the colon crypt epithelium and whether the crypt remains responsive to these cytokines during active damage and repair, remain unclear. Here, using laser-capture microdissection and primary colon organoid culture, we show that the cytokine milieu regulates the ability of the colonic crypt epithelium to participate in proinflammatory signaling. IFN-γ induces the TH1-recruiting, proinflammatory chemokine CXCL10/IP10 in primary murine intestinal crypt epithelium. CXCL10 was also induced in colonic organoids derived from mice with active, experimentally induced colitis, suggesting that the crypt can actively secrete CXCL10 in select cytokine environments during colitis. Colon expression of cxcl10 further increased during infectious and noninfectious colitis in Il17a−/−mice, demonstrating that IL-17A exerts a negative effect on CXCL10 in vivo. Furthermore, IL-17A directly antagonized CXCL10 production in ex vivo organoid cultures derived from healthy murine colons. Interestingly, direct antagonism of CXCL10 was not observed in organoids derived from colitic mouse colons bearing active lesions. These data, highlighting the complex interplay between the cytokine milieu and crypt epithelia, demonstrate proinflammatory chemokines can be induced within the colonic crypt and suggest the crypt remains responsive to cytokine modulation during inflammation.NEW & NOTEWORTHY Upon damage, the intestinal epithelium regenerates to restore barrier function. Here we observe that the local colonic cytokine milieu controls the production of procolitic chemokines within the crypt base and colon crypts remain responsive to cytokines during inflammation. IFN-γ promotes, while IL-17 antagonizes, CXCL10 production in healthy colonic crypts, while responses to cytokines differ in inflamed colon epithelium. These data reveal novel insight into colon crypt responses and inflammation-relevant alterations in signaling.

This work presents a new technique for 3D shape measurement of moving object in translational motion, which finds applications in online inspection, quality control, etc. A low-complexity 1D fast Fourier transform (FFT)-based spatial matching approach is devised to obtain accurate object displacement estimates, and it is combined with single shot fringe pattern prolometry (FPP) techniques to achieve high measurement performance with multiple captured images through coherent combining. The proposed technique overcomes some limitations of existing ones. Specifically, the placement of marks on object surface and synchronization between projector and camera are not needed, the velocity of the moving object is not required to be constant, and there is no restriction on the movement trajectory. Both simulation and experimental results demonstrate the effectiveness of the proposed technique.