Search

The integration of heterogeneous and unstructured data in Industry 4.0, poses a significant challenge, particularly with advanced manufacturing techniques. To address this issue, Knowledge Graphs (KGs) have emerged as a pivotal technology, yet their deployment often encounters the problem of incompletion due to data diversity and diverse storage formats. This study tackles the challenge of KG completion by applying and evaluating state-of-the-art KG embedding models-ComplEx, DistMult, TransE, ConvKB, and ConvE-within a football manufacturing production line context. Our analysis employs two principal metrics of Mean Reciprocal Rank (MRR) and Hits@N (Hits@10, Hits@3, and Hits@1) to comprehensively assess model performance. Our findings reveal that TransE significantly outperforms its counterparts, achieving an average accuracy of 91%, closely followed by ComplEx and DistMult with accuracies of 87% and 84%, respectively. Conversely, ConvKB and ConvE exhibit lower performance levels, with accuracy values of 79% and 76%. Through rigorous statistical testing, including t-tests, meaningful differences in MRR values across the models have been observed, with TransE leading in MRR and ConvE at the lower end of the spectrum. Our research not only sheds light on the efficacy of various KG embedding models in managing tree-like structured datasets within the manufacturing domain but also offers insights into optimising KGs for improved integration and analysis of data in production lines. These contributions are valuable both from academic research in KG completion and industrial practices aiming to enhance production efficiency and data coherence in advanced manufacturing settings.

High-precision geometric registration has been widely used in remote-sensing applications, such as image fusion. In practice, it is found that the geometric deformation (GD) of the bidirectional whisk-broom images, namely forward- and backward-scanning images, is not evenly distributed. Additionally, the deformation in some special areas exhibits significant nonlinear characteristics, which ultimately makes the traditional linear homography matrix inoperative in describing the GD of the whole image. Meanwhile, it is difficult to establish an elastic deformation model for the whole image due to the large swath and complex GD. To solve the problems above, we proposed a thin-plate-spline-based registration method for correcting the nonlinear deformation of large-size bidirectional whisk-broom images through scene classification. First, a classification method based on histogram Gaussian distribution fitting is proposed to divide the image into sea and land blocks of the same size. And the adjacent blocks must have overlapping areas. Second, the thin-plate-spline transformation is carried out to correct the local nonlinear deformation of the land blocks. Finally, the registered blocks are stitched by a gradual image fusion method to obtain the fully registered images. Our proposed method can correct the local nonlinear GD of bidirectional whisk-broom images of sustainable development science satellite 1 thermal infrared spectrometer (TIS), and the registration accuracy is better than 0.3 pixels, which is of great significance for the practical application of TIS thermal infrared image.

Gesture recognition has always been one of the important research directions in the field of computer vision. The dynamic gesture has the problems of complex backgrounds and many interference factors. The gesture recognition model based on deep learning usually has high computational cost and poor real-time performance. In addition, deep learning models are limited to recognizing existing categories in the training set and their performance largely depends on the amount of labeled data. To address the above problems, this paper presents a dynamic gesture recognition method named 3SCKI based on a three-stream coordinate attention (CA) network, knowledge distillation, and image-text contrastive learning. Specifically, 1) CA is utilized for feature fusion to make the model focus more on target gestures and reduce background interference, 2) traditional knowledge distillation loss is improved to reduce the amount of calculation and improve the real-time performance. Specifically, the guidance function is added to make the student network only learn the classification probability correctly identified by the teacher network, and 3) multi-granularity context prompt template integration method is proposed to construct an improved CLIP visual language model MG-CLIP. It aligns text and visual concepts from the image level to the object level to the part level. Through comparative learning of image features and text features, gesture classification is performed, enabling the model to identify image categories that have not appeared during the training phase. The proposed method is evaluated on the ChaLearn LAP large-scale isolated gesture dataset (IsoGD). The results show that our proposed method can obtain recognition rates of 65.87% on the validation set of IsoGD. For single mode data, 3SCKI can obtain the state-of-the-art recognition accuracy on RGB, Depth, and Optical Flow data (61.22%, 58.84%, and 50.30% of the validation set of IsoGD, respectively).

The growing complexity and interdisciplinary nature of Materials Science research demand efficient data management and exchange through structured knowledge representation. Domain-Level Ontologies (DLOs) for Materials Science have emerged as a valuable tool for describing materials properties, processes, and structures, enabling effective data integration, interoperability, and knowledge discovery. However, the harmonization of DLOs, and, more generally, the establishment of fully interoperable multi-level ecosystems, remains a challenge due to various factors, including the diverse landscape of existing ontologies. This work provides, for the first time in literature, a comprehensive overview of the state-of-the-art of DLOs for Materials Science, reviewing more than 40 DLOs and highlighting their main features and purposes. Furthermore, an alignment methodology including both manual and automated steps, making use of Top-Level Ontologies’ (TLO) capability of promoting interoperability, and revolving around the engineering of FAIR standalone entities acting as minimal data pipelines (“bridge concepts”), is presented. A proof of concept is also provided. The primary aspiration of this undertaking is to make a meaningful contribution towards the establishment of a unified ontology framework for Materials Science, facilitating more effective data integration and fostering interoperability across Materials Science subdomains.

Measured fractional vegetation cover (FVC) on the ground is very important for validation of the remote sensing products and algorithms. However, because of the influence of some factors such as the angle of illumination and vegetation density, the existence of vegetation shadows limits the accuracy of FVC estimation. This article proposes a deep learning method to reduce the FVC estimation error based on high dynamic range (HDR) images with vegetation shadows (HDR REC-DL method). The HDR REC-DL method can accurately extract FVC from HDR images with complex texture information on vegetation shadows. This method is based on the U-Net convolutional network structure for semantic segmentation of images containing vegetation shadows, and the segmentation results are less affected by vegetation types. Results from the HDR REC-DL method were highly similar to the vegetation segmentation results from visual interpretation. Values of the kappa coefficient, F1 score (F1), recall, and mean intersection over union of the HDR REC-DL method were 0.926, 0.942, 0.924, 0.916 for sunny weather and 0.903, 0.974, 0.983, and 0.895 for cloudy weather, respectively. Compared with the vegetation segmentation accuracy of the shadow-resistant algorithm, the HDR REC-DL method increases the kappa coefficient, F1, and mIOU by 21%, 16%, and 29% for sunny weather, and by 11.1%, 3.6%, and 10.3% for cloudy weather, respectively. The HDR REC-DL method provides a novel method for accurately estimating FVC from images containing vegetation shadows.

Land surface temperature (LST) data in the thermal infrared (TIR) band measured by the moderate-resolution imaging spectroradiometer (MODIS) instrument are critical for studying surface urban heat islands (SUHIs); however, these acquired TIR LST data are contaminated by clouds, so it is crucial to develop a method to generate cloud-free LST products. In this article, employing Tianjin as the research area, we combined the weather research and forecasting model with a random forest and a spatial optimization algorithm to propose a cloud-free MODIS-like model (WRFFM). The model can reconstruct cloud-free MODIS-like LSTs and SUHIs are studied. The spatial patterns of the WRFFM LSTs and the MODIS LSTs are consistent; the correlation coefficients in July and December range from 0.8 to 0.91 and 0.8 to 0.93, respectively, and the root mean square errors range from 0.5 to 3.8 K and 0.4 to 1.8 K, respectively, indicating that the modeled results are accurate. We use these WRFFM LSTs to study SUHIs and evaluate the deviations between the MODIS SUHIs and WRFFM SUHIs. When the proportion of clear-sky pixels is below 30%, the deviation is above 3 K, and when the proportion of clear-sky pixels is above 80%, the deviation is below 0.6 K. The results indicate that the developed model can be applied to improve the study of SUHIs and that the number of clear-sky pixels for a city is an important factor that affects the bias relative to the actual SUHI .

In optimizing the production of a metal mine, either the overall dynamic relations between technical indicators or the spatial distribution of the ore grade are usually considered, but few studies have considered both factors together. These two factors in combination have a greater effect on the optimization of mine production in terms of economic benefit and resource utilization than they do individually. We proposed an overall dynamic optimization model of technical indicators of metal mine production that considers the spatial distribution of the ore grade to better optimize the technical indicators and improve sustainable development of mineral resources. We incorporated an adaptive mutation strategy and adaptive control parameters into a differential evolution algorithm (AADE) in order to overcome the drawbacks of the differential evolution algorithm in solving this optimization model. The adaptive mutation strategy and adaptive control parameters were used to increase the rate of convergence and improve the search for a global maximum. To assess the performance of AADE, we used a real case and four test functions (the Sphere, Griewank, Rastrigin and Rosenbrock functions) in tests that compared AADE with a standard genetic algorithm, a standard differential evolution algorithm and the recently developed adaptive differential evolution algorithm. The results indicate that the optimization model we created is better aligned with mine production processes than current optimization models. In optimizing the technical indicators of metal mine production to maximize economic benefits, AADE performed significantly better than the other three algorithms tested in terms of convergence rate and global search ability.

This study evaluated the non-thermal effects of microwave on bovine lens by measuring the S-parameters induced by the dielectric coefficient variation in bovine lens. A closed, integrated biological experiment system was designed to expose bovine lens to microwave radiation quantitatively, and S-parameters were measured. Morphological evaluations were also performed during microwave radiation to evaluate the reaction mechanism of the non-thermal effects of microwave on the bovine lens. The results showed that the bovine lens became gradually cloudy and its refractive index changed significantly with increasing doses of microwave radiation. A comparison of experimental and control bovine lenses stained with hematoxylin and eosin revealed that the change in bovine lens might be a result of the nuclear fragmentation of lens epithelial cells and disorderly arrangement of lens fiber cells, which were induced by the vibration and friction of fiber tissues upset by microwave radiation. Further, the fluctuating decrements of 3.8 dB in S21 showed that the non-thermal effects of microwave could lead to the variations in the bovine lens dielectric coefficient. These findings suggest that the dielectric coefficient may quantitatively describe the physiological state of bovine lens. This study might be the valuable reference for the establishment of new electromagnetic radiation safety standard.

The analysis of influence spread in monoplex (single-layer) networks where all interactions are treated equally has been widely studied since the early 2000s. In complex networks, e.g., social networks, multi-agent networks, etc., the elements of networks interact in many ways, which results in multiplex (multiple-layer) networks. In this paper, we analyze the multiplexity-aware influence spread in two extended linear threshold models with two different protocols that explain how an agent updates its state combining signals from all layers. We then generalize the cohesiveness structural property to multiplex networks that explains why multiplexity facilitates or impedes a propagation. Based on this property, we present the optimal solutions by solving integer linear programming problems for influence minimization in multiplex linear threshold models.

Magneto-Acousto-Electric Tomography with coil detection, known as Magneto-Acousto-Electric Tomography with Magnetic Induction (MAET-MI), is a non-contact resistivity-based imaging method that employs a coil to detect the induced current generated by the ultrasound in biological tissue, which lie under a static magnetic field. To reconstruct an image of the tissue's conductivity, we propose a reciprocal model to describe the relationship between the inducted voltage of a coil and its conductivity. Previous work on the reciprocal theorem demonstrates that reconstructing conductivity using this method is effective. The forward and inverse problem are usually not verified both numerically and experimentally. In this paper, different approaches are adopted to calculate the forward and the inverse problems for verification of the reciprocal model. This verifies that the reconstruction method based on electrode detection can be applied to MAET-MI. This means that the inverse problem of MAET-MI can be transformed into an inverse source reconstruction of a wave equation based on the coil detection. In the forward problem, the moment method is employed to calculate the Radon transform and generate the ultrasonic signals. For the inverse problem, the filtered back projection method is chosen to reconstruct the ultrasound sources, which are related to the curl of the current density in the reciprocal process. Based on the curl of the current density in the reciprocal process, four sets of correlation coefficients of the original and reconstructed images' model are all greater than 90%. The uniform error criterion is obtained via multiple reconstructions and comparison of multiple models. The reciprocal model exhibits a good uniformity and stability when describing the actual physical process. It also provides additional effective ideas for solving the inverse problem quickly to reconstruct the ultrasonic sources, which is corresponding to the actual distribution of the conductivity.

Two-dimensional phase unwrapping (PU) is an essential step in interferometric synthetic aperture radar (InSAR) analysis. Although the $L^{0}$ -norm PU method is desired, it is nondeterministic polynomial (NP)-hard. Thus, many PU methods have been proposed to find near-optimal solutions of the $L^{0}$ -norm. As PU is an ill-posed problem, it is difficult to choose the method with the most accurate solution unless reference data are available. In this letter, we prove that the NP-hardness of the $L^{0}$ -norm is weak, suggesting that the $\alpha $ -approximation algorithm of the $L^{0}$ -norm can be devised, that is, the obtained near-optimal solutions can be within a factor of $\alpha $ of the $L^{0}$ -norm optimal value. The primary contribution of the proof is that the $\alpha $ value of each obtained PU solution can be considered as a new index to assess the PU performance without using reference data. The validity and effectiveness of the $\alpha $ -based index have been verified using simulated and acquired interferometric datasets and three PU methods. [ABSTRACT FROM AUTHOR]

The rapid progress of artificial intelligence expands its wide applicability in Internet of Things (IoT). Meanwhile, data insufficient and data source privacy are key supply chain challenges facing IoT especially in the healthcare industry. To address this problem in healthcare IoT, in this article, we propose a skin cancer detection model based on federated learning integrated with deep generation model. First, we employ dual generative adversarial networks to address the problem of insufficient data. In addition, to improve the quality of generated images, we synchronously optimize the sharpness of images, Frechet inception distance, image diversity, and loss using knee point-driven evolutionary algorithm (KnEA). Then, we protect patient information privacy by training federated skin cancer framework. Finally, we employ the ISIC 2018 dataset to test the performance of the proposed training model under different situations, including using identically distributed data, nonidentically distributed data, a sparse convolutional neural network, and a fully connected convolutional neural network. The experiment results demonstrate that the accuracy and area under the curve reach 91% and 88%, respectively. This model can help resolve problems of insufficient data in smart medicine of IoT and protect the privacy of user data while also providing an excellent detection rate. [ABSTRACT FROM AUTHOR]

Phase unwrapping (PU) is a key processing step in interferometric synthetic aperture radar (InSAR). To date, a number of skillful single-baseline (SB) and multibaseline (MB) PU methods exhibiting different advantages have been proposed. However, as the basic principles of SB and MB PUs are essentially different, it is difficult to effectively and systematically compare the performance of SB and MB PUs, despite the knowledge that this type of analysis is important for allowing the ever-increasing number of InSAR practitioners to choose a suitable approach for practical applications and to optimally plan future InSAR satellite missions. Recently, the framework of the two-stage programming approach (TSPA) was proposed, and this allows for the majority of the existing SB PU methods to be transplanted into the MB domain, to allow practitioners to feasibly and comprehensively compare SB and MB PU methodologies. In this study, the digital elevation model (DEM) reconstruction accuracy is compared between the classical SB PU methods and their corresponding TSPA-framework-based MB PU methods using the $L^{p}$ -norm model. Interestingly, we observed that although the number of PU residues in the MB case is larger than that in the SB case, the MB PU performance is better. The reason for this is that the type of MB residue is typically dipole, so the average length of the required MB branch-cut is shorter than that of SB. It is also demonstrated that the TSPA framework can effectively improve the PU accuracy of many existing SB PU methods when the number of input interferograms is sufficient. [ABSTRACT FROM AUTHOR]

The recent, sharp increase in the availability of interferometric data captured by different synthetic aperture radar (SAR) interferometry (InSAR) sensors poses a new scientific question that whether there is a processing framework that can combine these observations to obtain a more credible InSAR product (i.e., digital elevation model (DEM) and surface deformation estimation). In this article, we extend our previous two-stage programming-based multibaseline processing framework for combining the interferograms generated from disparate InSAR systems with different system parameters to enhance the InSAR performance at the signal processing stage. The proposed multisystem interferometric data fusion framework, abbreviated as TSDFF, includes three processing steps: multisystem interferogram registration, multisystem phase unwrapping, and absolute phase fusing. The advantage of TSDFF is that it can allow the data sets from different InSAR sensors to help each other to get rid of the limitation of the Itoh condition so that the application scope of each InSAR sensor will be effectively enlarged (e.g., measuring violent surface change or mountainous DEM). In addition, to quantitatively analyze the measurement bias robustness bound of TSDFF, the TSDFF-Fusion theorem is proposed, which offers significant application guidance for TSDFF at different noise levels. The real and simulated experimental results reveal the effectiveness of TSDFF for fusing the data sets from disparate InSAR systems. [ABSTRACT FROM AUTHOR]