Prediction of Drug-Target Interaction Using Dual-Network Integrated Logistic Matrix Factorization and Knowledge Graph Embedding.

Nowadays, drug-target interactions (DTIs) prediction is a fundamental part of drug repositioning. However, on the one hand, drug-target interactions prediction models usually consider drugs or targets information, which ignore prior knowledge between drugs and targets. On the other hand, models incorporating priori knowledge cannot make interactions prediction for under-studied drugs and targets. Hence, this article proposes a novel dual-network integrated logistic matrix factorization DTIs prediction scheme (Ro-DNILMF) via a knowledge graph embedding approach. This model adds prior knowledge as input data into the prediction model and inherits the advantages of the DNILMF model, which can predict under-studied drug-target interactions. Firstly, a knowledge graph embedding model based on relational rotation (RotatE) is trained to construct the interaction adjacency matrix and integrate prior knowledge. Secondly, a dual-network integrated logistic matrix factorization prediction model (DNILMF) is used to predict new drugs and targets. Finally, several experiments conducted on the public datasets are used to demonstrate that the proposed method outperforms the single base-line model and some mainstream methods on efficiency.

A novel stochastic resonance based deep residual network for fault diagnosis of rolling bearing system.

Rolling bearings constitute one of the most vital components in mechanical equipment, monitoring and diagnosing the condition of rolling bearings is essential to ensure safe operation. In actual production, the collected fault signals typically contain noise and cannot be accurately identified. In the paper, stochastic resonance (SR) is introduced into a spiking neural network (SNN) as a feature enhancement method for fault signals with varying noise intensities, combining deep learning with SR to enhance classification accuracy. The output signal-to-noise ratio(SNR) can be enhanced with the SR effect when the noise-affected fault signal input into neurons. Validation of the method is carried out through experiments on the CWRU dataset, achieving classification accuracy of 99.9%. In high-noise environments, with SNR equal to -8 dB, SRDNs achieve over 92% accuracy, exhibiting better robustness and adaptability.

MSL-Net: Sharp Feature Detection Network for 3D Point Clouds.

As a significant geometric feature of 3D point clouds, sharp features play an important role in shape analysis, 3D reconstruction, registration, localization, etc. Current sharp feature detection methods are still sensitive to the quality of the input point cloud, and the detection performance is affected by random noisy points and non-uniform densities. In this paper, using the prior knowledge of geometric features, we propose a Multi-scale Laplace Network (MSL-Net), a new deep-learning-based method based on an intrinsic neighbor shape descriptor, to detect sharp features from 3D point clouds. First, we establish a discrete intrinsic neighborhood of the point cloud based on the Laplacian graph, which reduces the error of local implicit surface estimation. Then, we design a new intrinsic shape descriptor based on the intrinsic neighborhood, combined with enhanced normal extraction and cosine-based field estimation function. Finally, we present the backbone of MSL-Net based on the intrinsic shape descriptor. Benefiting from the intrinsic neighborhood and shape descriptor, our MSL-Net has simple architecture and is capable of establishing accurate feature prediction that satisfies the manifold distribution while avoiding complex intrinsic metric calculations. Extensive experimental results demonstrate that with the multi-scale structure, MSL-Net has a strong analytical ability for local perturbations of point clouds. Compared with state-of-the-art methods, our MSL-Net is more robust and accurate.

Bitstream-based Perceptual Quality Assessment of Compressed 3D Point Clouds.

With the increasing demand of compressing and streaming 3D point clouds under constrained bandwidth, it has become ever more important to accurately and efficiently determine the quality of compressed point clouds, so as to assess and optimize the quality-of-experience (QoE) of end users. Here we make one of the first attempts developing a bitstream-based no-reference (NR) model for perceptual quality assessment of point clouds without resorting to full decoding of the compressed data stream. Specifically, we first establish a relationship between texture complexity and the bitrate and texture quantization parameters based on an empirical rate-distortion model. We then construct a texture distortion assessment model upon texture complexity and quantization parameters. By combining this texture distortion model with a geometric distortion model derived from Trisoup geometry encoding parameters, we obtain an overall bitstream-based NR point cloud quality model named streamPCQ. Experimental results show that the proposed streamPCQ model demonstrates highly competitive performance when compared with existing classic full-reference (FR) and reduced-reference (RR) point cloud quality assessment methods with a fraction of computational cost.

Deep Spiking Residual Shrinkage Network for Bearing Fault Diagnosis.

Bearing fault diagnosis of electrical equipment has been a popular research area in recent years because there are often some faults during continuous operation in production due to the harsh working environment. However, the traditional fault signal processing methods rely on highly expert experience, and some parameters are difficult to be optimized by machine-learning methods. Thus, the satisfactory recognition accuracy of fault diagnosis cannot be achieved in the above methods. In this article, a new model based on the spiking neural network (SNN) is proposed, which is called deep the spiking residual shrinkage network (DSRSN) for bearing fault diagnosis. In the model, attention mechanisms and soft thresholding are introduced to improve the recognition rate under a high-level noise background. The higher recognition accuracy is obtained in the proposed model which is tested on the fault signal dataset under different noise intensities. Meanwhile, the training time is about treble as fast as the training time of the artificial neural network, which is reflecting the high efficiency of SNN.

A no-Reference Stereoscopic Image Quality Assessment Network Based on Binocular Interaction and Fusion Mechanisms.

In contemporary society full of stereoscopic images, how to assess visual quality of 3D images has attracted an increasing attention in field of Stereoscopic Image Quality Assessment (SIQA). Compared with 2D-IQA, SIQA is more challenging because some complicated features of Human Visual System (HVS), such as binocular interaction and binocular fusion, must be considered. In this paper, considering both binocular interaction and fusion mechanisms of the HVS, a hierarchical no-reference stereoscopic image quality assessment network (StereoIF-Net) is proposed to simulate the whole quality perception of 3D visual signals in human cortex, including two key modules: BIM and BFM. In particular, Binocular Interaction Modules (BIMs) are constructed to simulate binocular interaction in V2-V5 visual cortex regions, in which a novel cross convolution is designed to explore the interaction details in each region. In the BIMs, different output channel numbers are designed to imitate various receptive fields in V2-V5. Furthermore, a Binocular Fusion Module (BFM) with automatic learned weights is proposed to model binocular fusion of the HVS in higher cortex layers. The verification experiments are conducted on the LIVE 3D, IVC and Waterloo-IVC SIQA databases and three indices including PLCC, SROCC and RMSE are employed to evaluate the assessment consistency between StereoIF-Net and the HVS. The proposed StereoIF-Net achieves almost the best results compared with advanced SIQA methods. Specifically, the metric values on LIVE 3D, IVC and WIVC-I are the best, and are the second-best on the WIVC-II.

New dual method for elastica regularization.

The Euler's elastica energy regularizer has been widely used in image processing and computer vision tasks. However, finding a fast and simple solver for the term remains challenging. In this paper, we propose a new dual method to simplify the solution. Classical fast solutions transform the complex optimization problem into simpler subproblems, but introduce many parameters and split operators in the process. Hence, we propose a new dual algorithm to maintain the constraint exactly, while using only one dual parameter to transform the problem into its alternate optimization form. The proposed dual method can be easily applied to level-set-based segmentation models that contain the Euler's elastic term. Lastly, we demonstrate the performance of the proposed method on both synthetic and real images in tasks image processing tasks, i.e. denoising, inpainting, and segmentation, as well as compare to the Augmented Lagrangian method (ALM) on the aforementioned tasks.

Iterative deep neural networks based on proximal gradient descent for image restoration.

The algorithm unfolding networks with explainability of algorithms and higher efficiency of Deep Neural Networks (DNN) have received considerable attention in solving ill-posed inverse problems. Under the algorithm unfolding network framework, we propose a novel end-to-end iterative deep neural network and its fast network for image restoration. The first one is designed making use of proximal gradient descent algorithm of variational models, which consists of denoiser and reconstruction sub-networks. The second one is its accelerated version with momentum factors. For sub-network of denoiser, we embed the Convolutional Block Attention Module (CBAM) in previous U-Net for adaptive feature refinement. Experiments on image denoising and deblurring demonstrate that competitive performances in quality and efficiency are gained by compared with several state-of-the-art networks for image restoration. Proposed unfolding DNN can be easily extended to solve other similar image restoration tasks, such as image super-resolution, image demosaicking, etc.

Multimodal Medical Supervised Image Fusion Method by CNN.

This article proposes a multimode medical image fusion with CNN and supervised learning, in order to solve the problem of practical medical diagnosis. It can implement different types of multimodal medical image fusion problems in batch processing mode and can effectively overcome the problem that traditional fusion problems that can only be solved by single and single image fusion. To a certain extent, it greatly improves the fusion effect, image detail clarity, and time efficiency in a new method. The experimental results indicate that the proposed method exhibits state-of-the-art fusion performance in terms of visual quality and a variety of quantitative evaluation criteria. Its medical diagnostic background is wide.

Craniofacial similarity analysis through sparse principal component analysis.

The computer-aided craniofacial reconstruction (CFR) technique has been widely used in the fields of criminal investigation, archaeology, anthropology and cosmetic surgery. The evaluation of craniofacial reconstruction results is important for improving the effect of craniofacial reconstruction. Here, we used the sparse principal component analysis (SPCA) method to evaluate the similarity between two sets of craniofacial data. Compared with principal component analysis (PCA), SPCA can effectively reduce the dimensionality and simultaneously produce sparse principal components with sparse loadings, thus making it easy to explain the results. The experimental results indicated that the evaluation results of PCA and SPCA are consistent to a large extent. To compare the inconsistent results, we performed a subjective test, which indicated that the result of SPCA is superior to that of PCA. Most importantly, SPCA can not only compare the similarity of two craniofacial datasets but also locate regions of high similarity, which is important for improving the craniofacial reconstruction effect. In addition, the areas or features that are important for craniofacial similarity measurements can be determined from a large amount of data. We conclude that the craniofacial contour is the most important factor in craniofacial similarity evaluation. This conclusion is consistent with the conclusions of psychological experiments on face recognition and our subjective test. The results may provide important guidance for three- or two-dimensional face similarity evaluation, analysis and face recognition.

CPU-GPU mixed implementation of virtual node method for real-time interactive cutting of deformable objects using OpenCL.

PURPOSE: Surgical simulators need to simulate interactive cutting of deformable objects in real time. The goal of this work was to design an interactive cutting algorithm that eliminates traditional cutting state classification and can work simultaneously with real-time GPU-accelerated deformation without affecting its numerical stability. METHODS: A modified virtual node method for cutting is proposed. Deformable object is modeled as a real tetrahedral mesh embedded in a virtual tetrahedral mesh, and the former is used for graphics rendering and collision, while the latter is used for deformation. Cutting algorithm first subdivides real tetrahedrons to eliminate all face and edge intersections, then splits faces, edges and vertices along cutting tool trajectory to form cut surfaces. Next virtual tetrahedrons containing more than one connected real tetrahedral fragments are duplicated, and connectivity between virtual tetrahedrons is updated. Finally, embedding relationship between real and virtual tetrahedral meshes is updated. Co-rotational linear finite element method is used for deformation. Cutting and collision are processed by CPU, while deformation is carried out by GPU using OpenCL. RESULTS: Efficiency of GPU-accelerated deformation algorithm was tested using block models with varying numbers of tetrahedrons. Effectiveness of our cutting algorithm under multiple cuts and self-intersecting cuts was tested using a block model and a cylinder model. Cutting of a more complex liver model was performed, and detailed performance characteristics of cutting, deformation and collision were measured and analyzed. CONCLUSIONS: Our cutting algorithm can produce continuous cut surfaces when traditional minimal element creation algorithm fails. Our GPU-accelerated deformation algorithm remains stable with constant time step under multiple arbitrary cuts and works on both NVIDIA and AMD GPUs. GPU-CPU speed ratio can be as high as 10 for models with 80,000 tetrahedrons. Forty to sixty percent real-time performance and 100-200 Hz simulation rate are achieved for the liver model with 3,101 tetrahedrons. Major bottlenecks for simulation efficiency are cutting, collision processing and CPU-GPU data transfer. Future work needs to improve on these areas.

Active contour model coupling with higher order diffusion for medical image segmentation.

Active contour models are very popular in image segmentation. Different features such as mean gray and variance are selected for different purpose. But for image with intensity inhomogeneities, there are no features for segmentation using the active contour model. The images with intensity inhomogeneities often occurred in real world especially in medical images. To deal with the difficulties raised in image segmentation with intensity inhomogeneities, a new active contour model with higher-order diffusion method is proposed. With the addition of gradient and Laplace information, the active contour model can converge to the edge of the image even with the intensity inhomogeneities. Because of the introduction of Laplace information, the difference scheme becomes more difficult. To enhance the efficiency of the segmentation, the fast Split Bregman algorithm is designed for the segmentation implementation. The performance of our method is demonstrated through numerical experiments of some medical image segmentations with intensity inhomogeneities.

Blurred palmprint recognition based on stable-feature extraction using a Vese-Osher decomposition model.

As palmprints are captured using non-contact devices, image blur is inevitably generated because of the defocused status. This degrades the recognition performance of the system. To solve this problem, we propose a stable-feature extraction method based on a Vese-Osher (VO) decomposition model to recognize blurred palmprints effectively. A Gaussian defocus degradation model is first established to simulate image blur. With different degrees of blurring, stable features are found to exist in the image which can be investigated by analyzing the blur theoretically. Then, a VO decomposition model is used to obtain structure and texture layers of the blurred palmprint images. The structure layer is stable for different degrees of blurring (this is a theoretical conclusion that needs to be further proved via experiment). Next, an algorithm based on weighted robustness histogram of oriented gradients (WRHOG) is designed to extract the stable features from the structure layer of the blurred palmprint image. Finally, a normalized correlation coefficient is introduced to measure the similarity in the palmprint features. We also designed and performed a series of experiments to show the benefits of the proposed method. The experimental results are used to demonstrate the theoretical conclusion that the structure layer is stable for different blurring scales. The WRHOG method also proves to be an advanced and robust method of distinguishing blurred palmprints. The recognition results obtained using the proposed method and data from two palmprint databases (PolyU and Blurred-PolyU) are stable and superior in comparison to previous high-performance methods (the equal error rate is only 0.132%). In addition, the authentication time is less than 1.3 s, which is fast enough to meet real-time demands. Therefore, the proposed method is a feasible way of implementing blurred palmprint recognition.

Automatic identification and morphometry of optic nerve fibers in electron microscopy images.

The neuroanatomical morphology of the optic nerve is an important description for understanding different aspects like topological distribution of nerves. Manual identification and morphometry has been usually considered as tedious, time consuming, and susceptible to error. A method that automates the identification and analysis of axons from electron micrographic images is presented. First, using region growing approach binarizes the image by combining the feature information together with spatial information, and obtains a coarse classification between myelin and non-myelin pixels. Next, identifies the axon candidates by region labeling and remove false axons on the basis of the identification ruler. Then the connected myelin sheaths are separated from each other using the maximum gradient magnitude of the outer annulus. Finally, analyses the morphological data of fibers. The developed method has been tested on a number of optic nerve images and results were presented. Regional distributions of axon caliber were unimodal. The thickness of the myelin sheath was highly correlated with the fiber diameter; hence, myelin sheath width was also distributed in a unimodal manner.