Dimensional Kernel Research Articles

Physicochemical and nutritional quality of pigmented rice and bran: Influence of milling and cooking

Pigmented rice bran is under-researched for human nutrition. This study investigated milling and cooking effects on the physicochemical and nutritional quality of pigmented and non-pigmented rice. Raw rice properties, cooking behavior and nutritional in-vitro quality were evaluated. Samples were cooked following a “risotto” preparation method. Bran removal by milling reduced kernel dimension (15%) and hardness (8%). Pigmented rice had higher total phenolic content (TPC; >41%) and higher antioxidant activity (AOA; >24%) than Carnaroli but milling significantly (p ≤ 0.05) reduced them. The bran was fractionated and (re)added during rice cooking as a new strategy to preserve nutritional compounds at their best. Compared with brown rice, milled samples exhibited greater geometric expansion and shorter cooking times (>47%), confirming that removal of the outer layer facilitated water penetration. Cooking affected the TPC differently depending on the rice variety (e.g., -15.5% for Violet and +7.8% for Orange brown samples). The incorporation of rice bran during cooking improved the “risotto” color and its nutritional value (TPC > 53% and AOA > 60%). Results evidenced the impact of different processing methods on rice quality, suggesting the potential use of rice co-product (i.e., bran) as new ingredient for healthier and more sustainable foods.

Omni-dimensional dynamic convolution feature coordinate attention network for pneumonia classification

Pneumonia is a serious disease that can be fatal, particularly among children and the elderly. The accuracy of pneumonia diagnosis can be improved by combining artificial-intelligence technology with X-ray imaging. This study proposes X-ODFCANet, which addresses the issues of low accuracy and excessive parameters in existing deep-learning-based pneumonia-classification methods. This network incorporates a feature coordination attention module and an omni-dimensional dynamic convolution (ODConv) module, leveraging the residual module for feature extraction from X-ray images. The feature coordination attention module utilizes two one-dimensional feature encoding processes to aggregate feature information from different spatial directions. Additionally, the ODConv module extracts and fuses feature information in four dimensions: the spatial dimension of the convolution kernel, input and output channel quantities, and convolution kernel quantity. The experimental results demonstrate that the proposed method can effectively improve the accuracy of pneumonia classification, which is 3.77% higher than that of ResNet18. The model parameters are 4.45M, which was reduced by approximately 2.5 times. The code is available at https://github.com/limuni/X-ODFCANET.

MDC-RHT: Multi-Modal Medical Image Fusion via Multi-Dimensional Dynamic Convolution and Residual Hybrid Transformer.

The fusion of multi-modal medical images has great significance for comprehensive diagnosis and treatment. However, the large differences between the various modalities of medical images make multi-modal medical image fusion a great challenge. This paper proposes a novel multi-scale fusion network based on multi-dimensional dynamic convolution and residual hybrid transformer, which has better capability for feature extraction and context modeling and improves the fusion performance. Specifically, the proposed network exploits multi-dimensional dynamic convolution that introduces four attention mechanisms corresponding to four different dimensions of the convolutional kernel to extract more detailed information. Meanwhile, a residual hybrid transformer is designed, which activates more pixels to participate in the fusion process by channel attention, window attention, and overlapping cross attention, thereby strengthening the long-range dependence between different modes and enhancing the connection of global context information. A loss function, including perceptual loss and structural similarity loss, is designed, where the former enhances the visual reality and perceptual details of the fused image, and the latter enables the model to learn structural textures. The whole network adopts a multi-scale architecture and uses an unsupervised end-to-end method to realize multi-modal image fusion. Finally, our method is tested qualitatively and quantitatively on mainstream datasets. The fusion results indicate that our method achieves high scores in most quantitative indicators and satisfactory performance in visual qualitative analysis.

Intelligent speech elderly rehabilitation learning assistance system based on deep learning and sensor networks

Recently, deep learning has been proved to significantly improve the quality of speech recognition. Convolutional neural networks are often used in speech recognition tasks because of their special network structure and powerful learning function. In order to solve the problem that the traditional convolutional neural network can not reflect the one-dimensional basic attributes of speech signal, this paper proposes to set the number of frames for one dimension of convolution kernel, and use one-dimensional model and two-dimensional convolutional network model for speech recognition. By moving the convolution kernel of time axis and frequency band, it can adapt to the time change of speech signal and maintain the correlation between frequency bands to a great extent. At the same time, this paper also discusses the speech signal preprocessing, feature parameter extraction and regularization algorithm. Due to the lack of hospital resources, the lag of information technology, the poor ability of accompanying and other reasons, the current accompanying service for elderly rehabilitation can not meet the needs of elderly patients. The rapid development and wide use of information technology provide opportunities for the optimization of health services for the elderly. In view of the difficulty of word memory and interpretation in current rehabilitation courses, a VR intelligent teaching system based on intelligent voice technology is developed. The application results show that the system can effectively improve the ability of language expression and word writing. At present, the system of intelligent speech function has not been completed, and lacks speech synthesis function. The next research will focus on the use of speech synthesis technology, in order to realize the man-machine dialogue between people and the system, and show a more real training situation.

SCSONet: spatial-channel synergistic optimization net for skin lesion segmentation

In the field of computer-assisted medical diagnosis, developing medical image segmentation models that are both accurate and capable of real-time operation under limited computational resources is crucial. Particularly for skin disease image segmentation, the construction of such lightweight models must balance computational cost and segmentation efficiency, especially in environments with limited computing power, memory, and storage. This study proposes a new lightweight network designed specifically for skin disease image segmentation, aimed at significantly reducing the number of parameters and floating-point operations while ensuring segmentation performance. The proposed ConvStem module, with full-dimensional attention, learns complementary attention weights across all four dimensions of the convolution kernel, effectively enhancing the recognition of irregularly shaped lesion areas, reducing the model’s parameter count and computational burden, thus promoting model lightweighting and performance improvement. The SCF Block reduces feature redundancy through spatial and channel feature fusion, significantly lowering parameter count while improving segmentation results. This paper validates the effectiveness and robustness of the proposed SCSONet on two public skin lesion segmentation datasets, demonstrating its low computational resource requirements. https://github.com/Haoyu1Chen/SCSONet.

Ensemble dynamics and information flow deduction from whole-brain imaging data.

The recent advancements in large-scale activity imaging of neuronal ensembles offer valuable opportunities to comprehend the process involved in generating brain activity patterns and understanding how information is transmitted between neurons or neuronal ensembles. However, existing methodologies for extracting the underlying properties that generate overall dynamics are still limited. In this study, we applied previously unexplored methodologies to analyze time-lapse 3D imaging (4D imaging) data of head neurons of the nematode Caenorhabditis elegans. By combining time-delay embedding with the independent component analysis, we successfully decomposed whole-brain activities into a small number of component dynamics. Through the integration of results from multiple samples, we extracted common dynamics from neuronal activities that exhibit apparent divergence across different animals. Notably, while several components show common cooperativity across samples, some component pairs exhibited distinct relationships between individual samples. We further developed time series prediction models of synaptic communications. By combining dimension reduction using the general framework, gradient kernel dimension reduction, and probabilistic modeling, the overall relationships of neural activities were incorporated. By this approach, the stochastic but coordinated dynamics were reproduced in the simulated whole-brain neural network. We found that noise in the nervous system is crucial for generating realistic whole-brain dynamics. Furthermore, by evaluating synaptic interaction properties in the models, strong interactions within the core neural circuit, variable sensory transmission and importance of gap junctions were inferred. Virtual optogenetics can be also performed using the model. These analyses provide a solid foundation for understanding information flow in real neural networks.

The solvability of inhomogeneous boundary-value problems in Sobolev spaces

The aim of the paper is to develop a general theory of solvability of linear inhomogeneous boundary-value problems for systems of ordinary differential equations of arbitrary order in Sobolev spaces. Boundary conditions are allowed to be overdetermined or underdetermined. They may contain derivatives, of the unknown vector-valued function, whose integer or fractional orders exceed the order of the differential equation. Similar problems arise naturally in various applications. The theory introduces the notion of a rectangular number characteristic matrix of the problem. The index and Fredholm numbers of this matrix coincide, respectively, with the index and Fredholm numbers of the inhomogeneous boundary-value problem. Unlike the index, the Fredholm numbers (i.e., the dimensions of the problem kernel and co-kernel) are unstable even with respect to small (in the norm) finite-dimensional perturbations. We give examples in which the characteristic matrix can be explicitly found. We also prove a limit theorem for a sequence of characteristic matrices. Specifically, it follows from this theorem that the Fredholm numbers of the problems under investigation are semicontinuous in the strong operator topology. Such a property ceases to be valid in the general case.

PO-YOLOv5: A defect detection model for solenoid connector based on YOLOv5.

Solenoid connectors play important role in electronic stability system design, with the features of small size, low cost, fast response time and high reliability. The main production process challenge for solenoid connectors is the accurate detection of defects, which is closely related to safe driving. Both faultless and defective products have similar color and shape at the defect location, making proper inspection challenging. To address these issues, we proposed a defect detection model called PO-YOLOv5 to achieve accurate defect detection for solenoid connectors. First, an additional prediction head was added to enable the model to acquire more semantic information to detect larger-scale defective features. Second, we introduced dynamic convolution to learn complementary connections between the four dimensions of the convolution kernel by utilizing its multidimensional attention mechanism. Replacing conventional convolution with dynamic convolution enhances the detection accuracy of the model and reduces the inference time. Finally, we validated PO-YOLOv5 versus the state-of-the-art object detection methods on the same solenoid connectors dataset. Experiments revealed that our proposed approach exhibited higher accuracy. The mAP (mean Average Precision) result of PO-YOLOv5 was found to be about 90.1%. Compared with the original YOLOv5, PO-YOLOv5 exhibited improved precision by about 3%.

A New Assistance Navigation Method for Substation Inspection Robots to Safely Cross Grass Areas.

With the development of intelligent substations, inspection robots are widely used to ensure the safe and stable operation of substations. Due to the prevalence of grass around the substation in the external environment, the inspection robot will be affected by grass when performing the inspection task, which can easily lead to the interruption of the inspection task. At present, inspection robots based on LiDAR sensors regard grass as hard obstacles such as stones, resulting in interruption of inspection tasks and decreased inspection efficiency. Moreover, there are inaccurate multiple object-detection boxes in grass recognition. To address these issues, this paper proposes a new assistance navigation method for substation inspection robots to cross grass areas safely. First, an assistant navigation algorithm is designed to enable the substation inspection robot to recognize grass and to cross the grass obstacles on the route of movement to continue the inspection work. Second, a three-layer convolutional structure of the Faster-RCNN network in the assistant navigation algorithm is improved instead of the original full connection structure for optimizing the object-detection boxes. Finally, compared with several Faster-RCNN networks with different convolutional kernel dimensions, the experimental results show that at the convolutional kernel dimension of 1024, the proposed method in this paper improves the mAP by 4.13% and the mAP is 91.25% at IoU threshold 0.5 in the range of IoU thresholds from 0.5 to 0.9 with respect to the basic network. In addition, the assistant navigation algorithm designed in this paper fuses the ultrasonic radar signals with the object recognition results and then performs the safety judgment to make the inspection robot safely cross the grass area, which improves the inspection efficiency.

SOLVABILITY OF FRACTIONAL FUNCTIONAL BOUNDARY-VALUE PROBLEMS WITH P-LAPLACIAN OPERATOR ON A HALF-LINE AT RESONANCE

This paper aims to consider the existence of solutions for p-Laplacian functional boundary-value problems at resonance on a half-line with two dimensional kernel. By employing some operators which satisfies suitable conditions and the Re and Gen extension of coincidence degree theory, a new result on the existence of solutions is acquired.

Travelling waves for Maxwell’s equations in nonlinear and nonsymmetric media

We look for travelling wave fields E(x,y,z,t)=U(x,y)cos(kz+ωt)+U~(x,y)sin(kz+ωt),(x,y,z)∈R3,t∈R\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\begin{aligned} E(x,y,z,t)&= U(x,y) \\cos (kz+\\omega t)+ {\\widetilde{U}}(x,y)\\sin (kz+\\omega t),\\\\&\\quad (x,y,z)\\in \\mathbb {R}^3,\\, t\\in \\mathbb {R} \\end{aligned}$$\\end{document}satisfying Maxwell’s equations in a nonlinear medium which is not necessarily cylindrically symmetric. The nonlinearity of the medium enters Maxwell’s equations by postulating a nonlinear material law D=varepsilon E+chi (x,y, langle |E|^2rangle )E between the electric field E, its time averaged intensity langle |E|^2rangle and the electric displacement field D. We derive a new semilinear elliptic problem for the profiles U,{widetilde{U}}:mathbb {R}^2rightarrow mathbb {R}^3Lu-V(x,y)u=f(x,y,u)withu=UU~,for(x,y)∈R2,\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\begin{aligned} Lu-V(x,y)u=f(x,y,u)\\quad \\hbox {with }u=\\begin{pmatrix} U \\\\ {\\widetilde{U}} \\end{pmatrix}, \\hbox { for }(x,y)\\in \\mathbb {R}^2, \\end{aligned}$$\\end{document}where f(x,y,u)=omega ^2chi (x,y, |u|^2)u. Solving this equation we can obtain exact travelling wave solutions of the underlying nonlinear Maxwell equations. We are able to deal with super quadratic and subcritical focusing effects, e.g. in the Kerr-like materials with the nonlinear susceptibility of the form chi (x,y,langle |E^2|rangle E) = chi ^{(3)}(x,y)langle |E|^2rangle E. A variational approach is presented for the semilinear problem. The energy functional associated with the equation is strongly indefinite, since L contains an infinite dimensional kernel. The methods developed in this paper may be applicable to other strongly indefinite elliptic problems and other nonlinear phenomena.

Molecular mapping of QTLs for grain dimension traits in Basmati rice.

Basmati rice is known for its extra-long slender grains, exceptional kernel dimensions after cooking, high volume expansion, and strong aroma. Developing high yielding Basmati rice varieties with good cooking quality is a gigantic task. Therefore, identifying the genomic regions governing the grain and cooked kernel dimension traits is of utmost importance for its use in marker-assisted breeding. Although several QTLs governing grain dimension traits have been reported, limited attempts have been made to map QTLs for grain and cooked kernel dimension traits of Basmati rice. In the current study, a population of recombinant inbred lines (RIL) was generated from a cross of Sonasal and Pusa Basmati 1121 (PB1121). In the RIL population, there was a significant positive correlation among the length (RRL: rough rice length, MRL: milled rice length, CKL: cooked kernel length) and breadth (RRB: rough rice breadth, MRB: milled rice breadth and CKB: cooked kernel breadth) of the related traits, while there was significant negative correlation between them. QTL mapping has led to the identification of four major genomic regions governing MRL and CKL. Two QTLs co-localize with the earlier reported major gene GS3 and a QTL qGRL7.1, while the remaining two QTLs viz., qCKL3.2 (qMRL3.2) and qCKL4.1 (qMRL4.1) were novel. The QTL qCKL3.2 has been bracketed to a genomic region of 0.78 Mb between the markers RM15247 and RM15281. Annotation of this region identified 18 gene models, of which the genes predicted to encode pentatricopeptides and brassinosteroid insensitive 1-associated receptor kinase 1 precursor may be the putative candidate genes. Furthermore, we identified a novel QTL qKER2.1 governing kernel elongation ratio (KER) in Basmati rice.

Optimization of chamfer masks using Farey sequences and kernel dimensionality

Farey sequences have captured the attention of several researchers because of their wide applications in polygonal approximation, generation of Ford circles, and shape analysis. In this work, we extend the applications of these sequences to optimize chamfer masks for computation of distance maps in images. Compared with previous methods, the proposed method can more effectively generate optimal weights from larger chamfer masks without considering multiple and rather complex defining variables of the masks. Furthermore, our work demonstrates the relationship between size of the chamfer kernel, Farey sequence, and optimal weights of the chamfer mask. This interesting relationship, which may be useful in various image processing and computer vision tasks, has never been revealed by any other previous study. Results from the current research may advance our understanding on the applications of Farey sequences in computational geometry and vision-related tasks. To allow reproducibility of the results, implementation codes and datasets can be accessed in the public repository at https://www.mathworks.com/matlabcentral/fileexchange/71652-optimization-of-chamfer-masks.

Sufficient dimension reduction for average causal effect estimation

A large number of covariates can have a negative impact on the quality of causal effect estimation since confounding adjustment becomes unreliable when the number of covariates is large relative to the number of samples. Propensity score is a common way to deal with a large covariate set, but the accuracy of propensity score estimation (normally done by logistic regression) is also challenged by the large number of covariates. In this paper, we prove that a large covariate set can be reduced to a lower dimensional representation which captures the complete information for adjustment in causal effect estimation. The theoretical result enables effective data-driven algorithms for causal effect estimation. Supported by the result, we develop an algorithm that employs a supervised kernel dimension reduction method to learn a lower dimensional representation from the original covariate space, and then utilises nearest neighbour matching in the reduced covariate space to impute the counterfactual outcomes to avoid the large sized covariate set problem. The proposed algorithm is evaluated on two semisynthetic and three real-world datasets and the results show the effectiveness of the proposed algorithm.

Efficient blood vessel segmentation from color fundus image using deep neural network

Blood vessel segmentation is an essential element of automatic retinal disease screening systems. In particular, retinal blood vessel analysis from fundus image is vital in the identification and diagnosis of cardiovascular and ophthalmological diseases (Ex: Diabetic Retinopathy, Macular degeneration, Retinal Pigmentosa, Macular Edema, and various stages of Glaucoma, etc). Wherefore, the diagnosis of these diseases by automatic vessel segmentation has become essential, especially in disclosure of premature prognosis of vision condition. In general, blood vessel extraction is divided into vessel tracking and pixel classification. In vessel tracking a vasculature model is expanded from a seed point. In pixel classification, the classifier classifies the pixels as either a vessel or background pixel, which is demonstrated in the proposed architecture. In this paper, deep learning based 19 layer U-Net architecture is proposed for the accurate and efficient segmentation of blood vessels. Prior to segmentation, a pre-processing block of AlexNet architecture is introduced for the classification of high-quality images from the experimented databases. This pre-classification stage helps in efficiently picking high-quality images determined by clarity, field definition, and sharpness. AlexNet classification is pivotal in enhancing the overall performance of the system by segmenting fine and tiny blood vessels. The proposed U-Net architecture has an encoder-decoder framework with 9 and 5 convolutional layers in each respectively. In order to boost the efficiency of the network as well as to reduce training and testing time, a proper choice of kernel dimension and number of filters are necessary. Our architecture was investigated on popular databases such as DRIVE, ARIA_d and MESSIDOR and various performance measures (accuracy, sensitivity, specificity, sensibility, Dice coefficient, and Jaccard coefficient) have been computed along with the Receiver Operating Characteristics. It is observed that the accuracy for DRIVE, ARIA_d and MESSIDOR are 90.60%, 87.60% and 83.42%, respectively. Area under curve in Receiver Operating Characteristics plot is found to be 98.54%, 93.28% and 88.18%, for DRIVE, ARIA_d and MESSIDOR databases, respectively. Results with the proposed architecture show remarkable improvement in the performance metrics for blood vessel segmentation.

Spatial-temporal Dynamic Forecasting of EVs Charging Load Based on DCC-2D

The charging load of electric vehicles (EVs) has a strong spatiotemporal randomness. Predicting the dynamic spatiotemporal distribution of the charging load of EVs is of great significance for the grid to cope with the access of large-scale EVs. Existing studies lack a prediction model that can accurately describe the dual dynamic changes of EVs charging the load time and space. Therefore, a spatial-temporal dynamic load forecasting model, dilated causal convolution-2D neural network (DCC-2D), is proposed. First, a hole factor is added to the time dimension of the three-dimensional convolutional convolution kernel to form a two-dimensional hole convolution layer so that the model can learn the spatial dimension information. The entire network is then formed by stacking the layers, ensuring that the network can accept long-term historical input, enabling the model to learn time dimension information. The model is simulated with the actual data of the charging pile load in a certain area and compared with the ConvLSTM model. The results prove the validity of the proposed prediction model.

Index-like theorem for massless fermions in spherically symmetric monopole backgrounds

In this paper we study massless fermions coupled to spherically symmetric SU(N) monopoles without Yukawa couplings between the Higgs and fermion fields. The corresponding Dirac operator is not Fredholm and the associated eigenfunctions are not L2-normalizable. Here we derive a formula for the dimension of the plane-wave normalizable kernel of such a Dirac operator for fermions of any representation of SU(N) in the presence of any spherically symmetric monopole background. Notably, our results also apply to fermions coupled to monopoles that preserve non-abelian gauge symmetry.

Enhancing Collaborative and Geometric Multi-Kernel Learning Using Deep Neural Network

This research proposes a method called enhanced collaborative and geometric multi-kernel learning (E-CGMKL) that can enhance the CGMKL algorithm which deals with multi-class classification problems with non-linear data distributions. CGMKL combines multiple kernel learning with softmax function using the framework of multi empirical kernel learning (MEKL) in which empirical kernel mapping (EKM) provides explicit feature construction in the high dimensional kernel space. CGMKL ensures the consistent output of samples across kernel spaces and minimizes the within-class distance to highlight geometric features of multiple classes. However, the kernels constructed by CGMKL do not have any explicit relationship among them and try to construct high dimensional feature representations independently from each other. This could be disadvantageous for learning on datasets with complex hidden structures. To overcome this limitation, E-CGMKL constructs kernel spaces from hidden layers of trained deep neural networks (DNN). Due to the nature of the DNN architecture, these kernel spaces not only provide multiple feature representations but also inherit the compositional hierarchy of the hidden layers, which might be beneficial for enhancing the predictive performance of the CGMKL algorithm on complex data with natural hierarchical structures, for example, image data. Furthermore, our proposed scheme handles image data by constructing kernel spaces from a convolutional neural network (CNN). Considering the effectiveness of CNN architecture on image data, these kernel spaces provide a major advantage over the CGMKL algorithm which does not exploit the CNN architecture for constructing kernel spaces from image data. Additionally, outputs of hidden layers directly provide features for kernel spaces and unlike CGMKL, do not require an approximate MEKL framework. E-CGMKL combines the consistency and geometry preserving aspects of CGMKL with the compositional hierarchy of kernel spaces extracted from DNN hidden layers to enhance the predictive performance of CGMKL significantly. The experimental results on various data sets demonstrate the superior performance of the E-CGMKL algorithm compared to other competing methods including the benchmark CGMKL.

Minimal σ-field for flexible sufficient dimension reduction

Sufficient Dimension Reduction (SDR) becomes an important tool for mitigating the curse of dimensionality in high dimensional regression analysis. Recently, Flexible SDR (FSDR) has been proposed to extend SDR by finding lower dimensional projections of transformed explanatory variables. The dimensions of the projections however cannot fully represent the extent of data reduction FSDR can achieve. As a consequence, optimality and other theoretical properties of FSDR are currently not well understood. In this article, we propose to use the σ-field associated with the projections, together with their dimensions to fully characterize FSDR, and refer to the σ-field as the FSDR σ-field. We further introduce the concept of minimal FSDR σ-field and consider FSDR projections with the minimal σ-field optimal. Under some mild conditions, we show that the minimal FSDR σ-field exists, attaining the lowest dimensionality at the same time. To estimate the minimal FSDR σ-field, we propose a two-stage procedure called the Generalized Kernel Dimension Reduction (GKDR) method and partially establish its consistency property under weak conditions. Extensive simulation experiments demonstrate that the GKDR method can effectively find the minimal FSDR σ-field and outperform other existing methods. The application of GKDR to a real life air pollution data set sheds new light on the connections between atmospheric conditions and air quality.

Hybrid dilated convolutional neural network for solving electromagnetic inverse scattering problems

In this article, a novel machine learning network is presented for solving 2-D electromagnetic (EM) inverse scattering problems (ISPs). The conventional approach of solving ISPs may suffer some difficulties such as the intrinsic strong nonlinearity, ill-conditioned problem, and high computational cost, especially when the scatterer is electrically large or contains high dielectric contrast. In order to solve the above problems, a novel hybrid dilated convolutional neural network (HDCNN) is proposed, which integrates with the advantage of the dilated convolution operation and the downsampling-upsampling (DSUS) algorithm. On one hand, the dilated convolution can expand the reception field valiantly and reduce the depth of neural network without extra computational cost. On the other hand, the DSUS algorithm can effectively overcome the irrelevance of the long-range data, when the dilated convolution is utilized. In addition, to avoid gridding problems, the HDCNN contains three dilated convolution layers with a different dilation rate, instead of adopting the same dimensional convolution kernel. Meanwhile, the proposed HDCNN could directly address the obtained scattering electric field without extra pretreatment. After executing the HDCNN process, EM parameters, for example, the relative permittivity, can be fast reconstructed. The efficiency and validity of the network can be demonstrated by four numerical results.