Conventional Neural Network Approach Research Articles

Transfer-Learning-Based Human Activity Recognition Using Antenna Array

Due to its low cost and privacy protection, Channel-State-Information (CSI)-based activity detection has gained interest recently. However, to achieve high accuracy, which is challenging in practice, a significant number of training samples are required. To address the issues of the small sample size and cross-scenario in neural network training, this paper proposes a WiFi human activity-recognition system based on transfer learning using an antenna array: Wi-AR. First, the Intel5300 network card collects CSI signal measurements through an antenna array and processes them with a low-pass filter to reduce noise. Then, a threshold-based sliding window method is applied to extract the signal of independent activities, which is further transformed into time–frequency diagrams. Finally, the produced diagrams are used as input to a pretrained ResNet18 to recognize human activities. The proposed Wi-AR was evaluated using a dataset collected in three different room layouts. The testing results showed that the suggested Wi-AR recognizes human activities with a consistent accuracy of about 94%, outperforming the other conventional convolutional neural network approach.

NeuronAlg: An Innovative Neuronal Computational Model for Immunofluorescence Image Segmentation

Background: Image analysis applications in digital pathology include various methods for segmenting regions of interest. Their identification is one of the most complex steps and therefore of great interest for the study of robust methods that do not necessarily rely on a machine learning (ML) approach. Method: A fully automatic and optimized segmentation process for different datasets is a prerequisite for classifying and diagnosing indirect immunofluorescence (IIF) raw data. This study describes a deterministic computational neuroscience approach for identifying cells and nuclei. It is very different from the conventional neural network approaches but has an equivalent quantitative and qualitative performance, and it is also robust against adversative noise. The method is robust, based on formally correct functions, and does not suffer from having to be tuned on specific data sets. Results: This work demonstrates the robustness of the method against variability of parameters, such as image size, mode, and signal-to-noise ratio. We validated the method on three datasets (Neuroblastoma, NucleusSegData, and ISBI 2009 Dataset) using images annotated by independent medical doctors. Conclusions: The definition of deterministic and formally correct methods, from a functional and structural point of view, guarantees the achievement of optimized and functionally correct results. The excellent performance of our deterministic method (NeuronalAlg) in segmenting cells and nuclei from fluorescence images was measured with quantitative indicators and compared with those achieved by three published ML approaches.

Monadic Pavlovian associative learning in a backpropagation-free photonic network

Over a century ago, Ivan P. Pavlov, in a classic experiment, demonstrated how dogs can learn to associate a ringing bell with food, thereby causing a ring to result in salivation. Today, it is rare to find the use of Pavlovian type associative learning for artificial intelligence applications even though other learning concepts, in particular, backpropagation on artificial neural networks (ANNs), have flourished. However, training using the backpropagation method on “conventional” ANNs, especially in the form of modern deep neural networks, is computationally and energy intensive. Here, we experimentally demonstrate a form of backpropagation-free learning using a single (or monadic) associative hardware element. We realize this on an integrated photonic platform using phase-change materials combined with on-chip cascaded directional couplers. We then develop a scaled-up circuit network using our monadic Pavlovian photonic hardware that delivers a distinct machine learning framework based on single-element associations and, importantly, using backpropagation-free architectures to address general learning tasks. Our approach reduces the computational burden imposed by learning in conventional neural network approaches, thereby increasing speed while also offering a higher bandwidth inherent to our photonic implementation.

Improved Intra Prediction Algorithm for HEVC with Conventional and Convolutional Neural Network Approach

Video compression has evolved since last decade and has extended its employability to many applications such as high definition televisions, video conferencing etc. High definition videos have nearly two times of bit rate than standard definition videos. Hence, video compression is an important approach to obtain information with lesser data size. The paper proposes an improved approach for intra-prediction in HEVC with modified resolution enhancement approach and intra prediction is considered for all 35 modes. The obtained results shows that the similarity between input and output is nearly same and PSNR average is around 42. Further, a deep learning approach has also been incorporated and coding unit depth is predicted for intra prediction. The developed model has provided 85% accuracy maintaining the similar PSNR.

Attention Guided Lymph Node Malignancy Prediction in Head and Neck Cancer.

Accurate lymph node (LN) malignancy classification is essential for treatment target identification in head and neck cancer (HNC) radiation therapy. Given the constraints imposed by relatively small sample sizes in real-world medical applications, to classify LN malignancy status accurately, we proposed an attention-guided classification (AGC) scheme that (1) incorporates human knowledge (ie, LN contours) into model training to guide model's "learning" direction, alleviating the critical requirement of large training samples by deep learning approaches; and (2) does not require accurate delineation of LNs in the inference stage but can highlight the discriminative region nearby the LN, which is important for malignancy determination. In the proposed AGC scheme, there is an attention-guided convolutional neural network (agCNN) module, followed by a classification convolutional neural network (cCNN) module. The input of the proposed AGC scheme is a region of interest (ROI) containing the LN and its surrounding tissues. The agCNN is designed to find the discriminative region in the ROI, which outputs an activation map whose voxel values indicate the importance of the voxels in malignancy prediction. Through multiplying the activation map with the ROI, we obtain the input for the cCNN, which finally outputs the LN malignancy probability. To demonstrate the effectiveness of the proposed scheme, we performed experimental studies using positron emission tomography and contrast-enhanced computed tomography from 129 surgical HNC patients, including 791 LNs, with pathologic ground truth of malignancy status. To evaluate the performance, 5-folder cross validation was used. The sensitivity, specificity, accuracy, and area under the receiver operating characteristic (ROC) curve values obtained by the proposed AGC scheme were 0.91, 0.93, 0.92, and 0.98, respectively, significantly outperforming conventional convolutional neural network and radiomics approaches at a significance level of .05 under a paired ROC comparison statistical test. We developed an AGC scheme that can highlight the discriminative region in an image for LN malignancy prediction, outperforming a conventional radiomics method that requires accurate segmentation and a standard convolutional neural network model without involving segmentation.

Surface Defect Inspection in Images Using Statistical Patches Fusion and Deeply Learned Features

Defect detection in images is a challenging task due to the existence of tiny and noisy patterns on surface images. To tackle this challenge, a defect detection approach is proposed in this paper using statistical data fusion. First, the proposed approach breaks a large image that contains multiple separate defects into smaller overlapping patches to detect the existence of defects in each patch, using the conventional convolutional neural network approach. Then, a statistical data fusion approach is proposed to maintain the spatial coherence of cracks in the image and aggregate the information extracted from overlapping patches to enhance the overall performance and robustness of the system. The proposed approach is evaluated using three benchmark datasets to demonstrate its superior performance in terms of both individual patch inspection and the whole image inspection.

Prediction of Porosity and Water Saturation Using Neural Networks in Shaly Sand Reservoirs, Western Deseret, Egypt

Petrophysical properties evaluation of shaly sandstone reservoirs is a challenging task in comparison to clean sand reservoirs. Logging derived porosity in shaly sands requires shale correction and Archie’s formula cannot be used in shaly sands for the determination of water saturation, therefore many water saturation models were proposed to get accurate water saturation of shaly sand reservoirs. In this paper, three water saturation models were used; two empirical models (Simandoux and total shale) and one theoretical model (effective medium model). Shale corrected density log was used in all models. The use of computer-generated algorithm, fuzzy log neural network is of increasing interest in the petroleum industry. This paper presents artificial neural network (ANN) as an effective tool for determining porosity and water saturation in shaly sand reservoir using well logging data. ANN technique utilizes the prevailing unknown nonlinear relationship in data between input logging data and output petrophysical parameters. Results of this work showed that ANN can be supplement or replacement of the existing conventional techniques to determine porosity and water saturation using empirical or theoretical water saturation models. Two neural networks were presented to determine porosity and water saturation using GR, resistivity and density logging data and adapted cut off for porosity and water saturation. Water saturation and porosity were determined using conventional techniques and neural network approach for two wells in a shaly sand reservoir. Neural network approach was trained for porosity and water saturation using the available well logging data. The predicted porosity and water saturation values have shown good matching with the core data in the two wells in comparison to the porosity and water saturation derived from the conventional techniques. This work showed that developed neural network (ANN) could provide an accurate porosity and water saturation in shaly sands reservoirs, it is subject to volume of available well logging data.

Prediction of stenosis behaviour in artery by neural network and multiple linear regressions.

Blood flow analysis in the artery is a paramount study in the field of arterial stenosis evaluation. Studies conducted so far have reported the analysis of blood flow parameters using different techniques, but the regression analysis is not adequately used. Artificial neural network is a nonlinear and nonparametric approach. It uses back-propagation algorithm for regression analysis, which is effective as compared to statistical model that requires a higher domain of statistics for prediction. In our manuscript, twofold analyses of data are done. First phase involves the determination of blood flow parameters using physiological flow pulse generator. The second phase includes regression modelling. The inputs to the model were axial length from stenosis, radial distance, inlet velocity, mean pressure, density, viscosity, time, and degree of blockage. Output included dependent variables in the form of output as mean velocity, root-mean-square (RMS) velocity, turbulent intensity, mean frequency, RMS frequency, frequency of turbulent intensity, gate time mean, gate time RMS. The temperature, density, and viscosity conditions were kept constant for various degrees of blockages. It was followed by regression analysis of variables using conventional statistical and neural network approach. The result shows that the neural network model is more appropriate, because value of percentage of response variation of dependent variable is almost approaching unity as compared to statistical analysis.

Redundancy Avoidance for Big Data in Data Centers: A Conventional Neural Network Approach

As the innovative data collection technologies are applying to every aspect of our society, the data volume is skyrocketing. Such phenomenon poses tremendous challenges to data centers with respect to enabling storage. In this paper, a hybrid-stream big data analytics model is proposed to perform multimedia big data analysis. This model contains four procedures, i.e., data pre-processing, data classification, data recognition and data load reduction. Specifically, an innovative multi-dimensional Convolution Neural Network (CNN) is proposed to assess the importance of each video frame. Thus, those unimportant frames can be dropped by a reliable decision-making algorithm. In order to ensure video quality, minimal correlation and minimal redundancy (MCMR) are combined to optimize the decision-making algorithm. Simulation results show that the amount of processed video is significantly reduced, and the quality of video is preserved due to the addition of MCMR. The simulation also proves that the proposed model performs steadily and is robust enough to scale up to accommodate the big data crush in data centers.

Object Detection in Ground-Penetrating Radar Images Using a Deep Convolutional Neural Network and Image Set Preparation by Migration

Ground-penetrating radar allows the acquisition of many images for investigation of the pavement interior and shallow geological structures. Accordingly, an efficient methodology of detecting objects, such as pipes, reinforcing steel bars, and internal voids, in ground-penetrating radar images is an emerging technology. In this paper, we propose using a deep convolutional neural network to detect characteristic hyperbolic signatures from embedded objects. As a first step, we developed a migration-based method to collect many training data and created 53510 categorized images. We then examined the accuracy of the deep convolutional neural network in detecting the signatures. The accuracy of the classification was 0.945 (94.5%)–0.979 (97.9%) when using several thousands of training images and was much better than the accuracy of the conventional neural network approach. Our results demonstrate the effectiveness of the deep convolutional neural network in detecting characteristic events in ground-penetrating radar images.

Neural network based optimization approach for energy demand prediction in smart grid

Energy usage and demand forecasting is an essential and complex task in real time implementation. Proper coordination is required between the consumer and power companies for monitoring, scheduling and operating the electrical devices without any damages. In this paper, we propose a novel neural network based optimization approach for energy demand prediction. Initially, the Conventional Neural Network (CNN) approach is employed to find the required energy demand prediction at the consumer end. Secondly, Neural Network based Genetic Algorithm (NNGA) and Neural Network based Particle Swarm Optimization (NNPSO) approaches are implemented where the weights of the neural network are automatically adjusted. Closer observation from the result reveals that the proposed NNGA approach performs better for short term load forecasting and proposed NNPSO is more suitable for the long term energy prediction. For the experimental results, real time data are taken from pecan street (Pecan Street Inc.). From the simulations, it can be concluded that the proposed optimization approach algorithm yield better results than the CNN approach in predicting the future energy demand. Further, the result reveals that it is possible to manage the demand and supply, planning of power grid and prediction of future energy requirement in the smart grid.

A computational intelligence approach to efficiently predicting review ratings in e-commerce

Sentiment analysis, also called opinion mining, is currently one of the most studied research fields which aims to analyse people's opinions. E-commerce websites allow users to share opinions about a product/service by providing textual reviews along with numerical ratings. These opinions greatly influence future consumer purchasing decisions. This paper introduces an innovative computational intelligence framework for efficiently predicting customer review ratings. The framework has been designed to deal with the dimensionality and noise which is typically apparent in large datasets containing customer reviews. The proposed framework integrates the techniques of Singular Value Decomposition (SVD) and dimensionality reduction, Fuzzy C-Means (FCM) and the Adaptive Neuro-Fuzzy Inference System (ANFIS). The performance of the proposed approach returned high accuracy and the results revealed that when large datasets are concerned, only a fraction of the data is needed for creating a system to predict the review ratings of textual reviews. Results from the experiments suggest that the proposed approach yields better prediction performance than other state-of-the-art rating predictors which are based on the conventional Artificial Neural Network, Fuzzy C-Means, and Support Vector Machine approaches. In addition, the proposed framework can be utilised for other classification and prediction tasks, and its neuro-fuzzy predictor module can be replaced by other classifiers.

Neural-Network Modeling for 3-D Substructures Based on Spatial EM-Field Coupling in Finite-Element Method

This paper presents a new neural-network method to describe the electromagnetic (EM) behavior at the interface between the substructures from an internally decomposed EM structure. A set of neural networks is used to represent the EM behavior of the substructure as seen from the interface. This allows EM coupling between substructures to be effectively represented. The method is developed in a finite-element environment. An EM transfer function matrix is formulated to produce training data, allowing neural networks to learn the spatial coupling between EM-field variables at various locations over the interface of the substructure. A new formulation is proposed where trained neural networks are integrated into the finite-element equation for efficient simulation of an overall EM structure. A technique is developed to allow the proposed model to be used with the mesh different from that in neural-network training. Examples show that the proposed method provides better accuracy than conventional neural-network approaches for modeling substructures from an internally decomposed EM problem. Using the proposed model also speeds up finite-element simulation.

Comparison of RBF and MLP neural networks to solve inverse kinematic problem for 6R serial robot by a fusion approach

In this paper, a fusion approach to determine inverse kinematics solutions of a six degree of freedom serial robot is proposed. The proposed approach makes use of radial basis function neural network for prediction of incremental joint angles which in turn are transformed into absolute joint angles with the assistance of forward kinematics relations. In this approach, forward kinematics relations of robot are used to obtain the data for training of neural network as well to estimate the deviation of predicted inverse kinematics solution from the desired solution. The effectiveness of the fusion process is shown by comparing the inverse kinematics solutions obtained for an end-effector of industrial robot moving along a specified path with the solutions obtained from conventional neural network approaches as well as iterative technique. The prominent features of the fusion process include the accurate prediction of inverse kinematics solutions with less computational time apart from the generation of training data for neural network with forward kinematics relations of the robot.

A hybrid neural network model for typhoon-rainfall forecasting

A hybrid neural network model is proposed in this paper to forecast the typhoon rainfall. Two different types of artificial neural networks, the self-organizing map (SOM) and the multilayer perceptron network (MLPN), are combined to develop the proposed model. In the proposed model, a data analysis technique is developed based on the SOM, which can perform cluster analysis and discrimination analysis in one step. The MLPN is used as the nonlinear regression technique to construct the relationship between the input and output data. First, the input data are analyzed using a SOM-based data analysis technique. Through the SOM-based data analysis technique, input data with different properties are first divided into distinct clusters, which can help the multivariate nonlinear regression of each cluster. Additionally, the topological relationships among data are discovered from which more insight into the typhoon-rainfall process can be revealed. Then, for each cluster, the individual relationship between the input and output data is constructed by a specific MLPN. For evaluating the forecasting performance of the proposed model, an application is conducted. The proposed model is applied to the Tanshui River Basin to forecast the typhoon rainfall. The results show that the proposed model can forecast more precisely than the model developed by the conventional neural network approach.

A Novel Strategy Using Factor Graphs and the Sum-Product Algorithm for Satellite Broadcast Scheduling Problems

This paper presents a low complexity algorithmic framework for finding a broadcasting schedule in a low-altitude satellite system, i. e., the satellite broadcast scheduling (SBS) problem, based on the recent modeling and computational methodology of factor graphs. Inspired by the huge success of the low density parity check (LDPC) codes in the field of error control coding, in this paper, we transform the SBS problem into an LDPC-like problem through a factor graph instead of using the conventional neural network approaches to solve the SBS problem. Based on a factor graph framework, the soft-information, describing the probability that each satellite will broadcast information to a terminal at a specific time slot, is exchanged among the local processing in the proposed framework via the sum-product algorithm to iteratively optimize the satellite broadcasting schedule. Numerical results show that the proposed approach not only can obtain optimal solution but also enjoys the low complexity suitable for integral-circuit implementation.

Neuro‐fuzzy structural classification of proteins for improved protein secondary structure prediction

Fourier transform infrared (FTIR) spectroscopy is a very flexible technique for characterization of protein secondary structure. Measurements can be carried out rapidly in a number of different environments based on only small quantities of proteins. For this technique to become more widely used for protein secondary structure characterization, however, further developments in methods to accurately quantify protein secondary structure are necessary. Here we propose a structural classification of proteins (SCOP) class specialized neural networks architecture combining an adaptive neuro-fuzzy inference system (ANFIS) with SCOP class specialized backpropagation neural networks for improved protein secondary structure prediction. Our study shows that proteins can be accurately classified into two main classes "all alpha proteins" and "all beta proteins" merely based on the amide I band maximum position of their FTIR spectra. ANFIS is employed to perform the classification task to demonstrate the potential of this architecture with moderately complex problems. Based on studies using a reference set of 17 proteins and an evaluation set of 4 proteins, improved predictions were achieved compared to a conventional neural network approach, where structure specialized neural networks are trained based on protein spectra of both "all alpha" and "all beta" proteins. The standard errors of prediction (SEPs) in % structure were improved by 4.05% for helix structure, by 5.91% for sheet structure, by 2.68% for turn structure, and by 2.15% for bend structure. For other structure, an increase of SEP by 2.43% was observed. Those results were confirmed by a "leave-one-out" run with the combined set of 21 FTIR spectra of proteins.

Urban traffic flow prediction using a fuzzy-neural approach

This paper develops a fuzzy-neural model (FNM) to predict the traffic flows in an urban street network, which has long been considered a major element in the responsive urban traffic control systems. The FNM consists of two modules: a gate network (GN) and an expert network (EN). The GN classifies the input data into a number of clusters using a fuzzy approach, and the EN specifies the input–output relationship as in a conventional neural network approach. While the GN groups traffic patterns of similar characteristics into clusters, the EN models the specific relationship within each cluster. An online rolling training procedure is proposed to train the FNM, which enhances its predictive power through adaptive adjustments of the model coefficients in response to the real-time traffic conditions. Both simulation and real observation data are used to demonstrative the effectiveness of the method.

Detection of delayed gastric emptying from electrogastrograms with support vector machine.

A recent study reported a conventional neural network (NN) approach for the noninvasive diagnosis of delayed gastric emptying from the cutaneous electrogastrograms. Using support vector machine, we show that this relatively new technique can be used for detection of delayed gastric emptying and is in fact able to outdo the conventional NN.

Adaptive reinforcement learning system for linearization control

A linearization scheme is proposed to demonstrate how a neural network scheme learns to linearize a system without any identification. The process occurs within an evaluator and a controller, which communicate with each other through reinforcement signals. From simulation results, the proposed learning scheme notably surpasses the conventional neural network approaches.