Search

Robotic arms have been widely used in industrial fields. However, researchers have seldom considered the factors affecting the actual factory environment. For example, when objects are conveyed in a factory, conveyor belts are often used to dynamically plan the overall production line. In addition, each object requires multiple checkpoints for repeated audits and inspections to ensure its quality. In this study, a vision-based robotic arm system equipped with multiple functionalities was developed. The development process consisted of three steps: detecting multiple dynamic objects, determining the size of each object, and identifying object defects. In the first step, You Only Look Once was used to detect multiple dynamic objects on a conveyor belt in real time. In the second step, the original image of the object was converted into a grayscale image, and the edge contour of the object was drawn using a Canny edge detection algorithm. Objects in the image were then rotated for vertical and horizontal projections, and then an artificial neural network (ANN) was used to calculate the size of each object. In the third step, a convolutional fuzzy neural network (CFNN) was used to identify object defects. This network was divided into an input layer, a convolution pooling layer, a feature fusion layer, a fuzzy layer, a regular layer, and a defuzzification layer. According to the experimental results, the standard error of the mean between the object size obtained by the ANN and the actual size was 0.009. In addition, the accuracy, recall, precision, and F1-score obtained by the CFNN in object defect detection were 0.9580, 0.9535, 0.9535, and 0.9535, respectively. Compared with other deep neural network models, such as AlexNet and LeNet, the proposed CFNN has fewer parameters and higher performance. [ABSTRACT FROM AUTHOR]

Spiking Neural Network (SNN) has been recognized as the third generation of neural networks. Conventionally, a SNN can be converted from a pre-trained Artificial Neural Network (ANN) with less computation and memory than training from scratch. But, these converted SNNs are vulnerable to adversarial attacks. Numerical experiments demonstrate that the SNN trained by optimizing the loss function will be more adversarial robust, but the theoretical analysis for the mechanism of robustness is lacking. In this paper, we provide a theoretical explanation by analyzing the expected risk function. Starting by modeling the stochastic process introduced by the Poisson encoder, we prove that there is a positive semidefinite regularizer. Perhaps surprisingly, this regularizer can make the gradients of the output with respect to input closer to zero, thus resulting in inherent robustness against adversarial attacks. Extensive experiments on the CIFAR10 and CIFAR100 datasets support our point of view. For example, we find that the sum of squares of the gradients of the converted SNNs is 13 ∼ 160 times that of the trained SNNs. And, the smaller the sum of the squares of the gradients, the smaller the degradation of accuracy under adversarial attack. • The adversarial robustness of SNNs based on rate encoder is explained mathematically. • A regularizer distinguishes the directly trained SNN from the converted SNN. • The coding length can also change the gradient of the input to affect the robustness. • We compare more types of the trained SNN with the converted SNN. [ABSTRACT FROM AUTHOR]

The precision of the machining tool in computer numerical control (CNC) machining is affected by several factors. For example, cutting parameters considerably affect machining accuracy and tool wear. Tool wear results in the manufacture of substandard products. Therefore, predicting tool wear is crucial in CNC machining. In this study, we proposed a backpropagation neural network (BPNN) to predict tool wear. In machine learning, backpropagation is a widely used algorithm for training artificial neural networks. The proposed BPNN considered the variation of tool wear with different cutting parameters, such as the spindle speed, feed, cutting depth, and cutting time. The experimental results revealed that the root mean square error of the BPNN prediction model was less than that of the linear regression prediction model. Furthermore, the proposed model achieved a coefficient of determination (R2) of 0.9964, which indicated that the BPNN model can accurately predict tool wear. [ABSTRACT FROM AUTHOR]

Deep learning has accomplished huge success in computer vision applications such as self-driving vehicles, facial recognition, and controlling robots. A growing need for deploying systems on resource-limited or resource-constrained environments such as smart cameras, autonomous vehicles, robots, smartphones, and smart wearable devices drives one of the current mainstream developments of convolutional neural networks: reducing model complexity but maintaining fine accuracy. In this study, the proposed efficient light convolutional neural network (ELNet) comprises three convolutional modules which perform ELNet using fewer computations, which is able to be implemented in resource-constrained hardware equipment. The classification task using CIFAR-10 and CIFAR-100 datasets was used to verify the model performance. According to the experimental results, ELNet reached 92.3% and 69%, respectively, in CIFAR-10 and CIFAR-100 datasets; moreover, ELNet effectively lowered the computational complexity and parameters required in comparison with other CNN architectures. [ABSTRACT FROM AUTHOR]

Skeleton-based hand gesture recognition is an active research topic in computer graphics and computer vision and has a wide range of applications in VR/AR and robotics. Although the spatial–temporal graph convolutional network has been successfully used in skeleton-based hand gesture recognition, these works often use a fixed spatial graph according to the hand skeleton tree or use a fixed graph on the temporal dimension, which may not be optimal for hand gesture recognition. In this paper, we propose a two-stream graph attention convolutional network with spatial–temporal attention for hand gesture recognition. We adopt pose stream and motion stream as the two input streams for our network. In pose stream, we use the joint in each frame as the input; In motion stream, we use the joint offsets between neighboring frames as the input. We propose a new temporal graph attention module to model the temporal dependency and also use a spatial graph attention module to construct dynamic skeleton graph. For each stream, we adopt graph convolutional network with spatial–temporal attention to extract the features. Then, we concatenate the feature of the pose stream and motion stream for gesture recognition. We achieve the competitive performance on the main hand gesture recognition benchmark datasets, which demonstrates the effectiveness of our method. [ABSTRACT FROM AUTHOR]

Lung cancer is one of the common causes of cancer deaths. Early detection and treatment of lung cancer is essential. However, the detection of lung cancer in patients produces many false positives. Therefore, increasing the accuracy of the classification of diagnosis or true detection by computed tomography (CT) is a difficult task. Solving this problem using intelligent and automated methods has become a hot research topic in recent years. Hence, we propose a 2D convolutional neural network (2D CNN) with Taguchi parametric optimization for automatically recognizing lung cancer from CT images. In the Taguchi method, 36 experiments and 8 control factors of mixed levels were selected to determine the optimum parameters of the 2D CNN architecture and improve the classification accuracy of lung cancer. The experimental results show that the average classification accuracy of the 2D CNN with Taguchi parameter optimization and the original 2D CNN in lung cancer recognition are 91.97% and 98.83% on the Lung Image Database Consortium and Image Database Resource Initiative (LIDC-IDRI) dataset, and 94.68% and 99.97% on the International Society for Optics and Photonics with the support of the American Association of Physicists in Medicine (SPIE-AAPM) dataset, respectively. The proposed method is 6.86% and 5.29% more accurate than the original 2D CNN on the two datasets, respectively, proving the superiority of proposed model. [ABSTRACT FROM AUTHOR]

Manufacturers of printed circuit boards (PCBs) typically use automated optical inspection (AOI) machines to test their PCBs. However, AOI machines employ conventional image-processing methods. If the integrated circuit (IC) components are not identical to the golden samples, then the AOI machine registers those IC components as flaws. Conventional image-processing methods cause misjudgments and increase the cost of manual reviews. Character-verification and image-classification systems are proposed in this paper for detecting misplaced, missing, and reversed-polarity parts. The regions of IC components can be identified on PCBs by using the contour border-detection method. Through the proposed convolutional neural network (CNN) structure and refinement mechanism, the characters can be successfully recognized. The image-classification system was applied only to images with blurry characters. Different CNN learning structures were used in both systems, and the refinement mechanism was used in both systems to improve the results. The proposed character-verification and image-classification methods achieved 98.84% and 99.48% passing rates, and the amount of required training time was less than that of other methods, demonstrating the proposed methods' greater effectiveness. [ABSTRACT FROM AUTHOR]

This paper presents a two-stage algorithm for automatic text detection and recognition. In the first stage, using a stroke width transform and an improved connected component, an edge analysis method detects a candidate character region. Subsequently, a text region is located by filtering and linking characters with similar font sizes and colors. For the second stage, a histogram of oriented gradient is employed as a feature descriptor, and a neural network classifier is built with dynamic-group-based hybrid particle swarm optimization (DGHPSO) for character recognition. In DGHPSO, each group's threshold value of similarity depends on the threshold values of fitness and distance. In addition, a local search algorithm is used to improve the search for a global optimum. The proposed algorithm was experimentally validated; it outperformed a number of recently published studies in terms of the text recognition rate when tested on the ICDAR 2003 database and the Street View Text database. [ABSTRACT FROM AUTHOR]

When photographs are being taken in an outdoor environment, the medium in air will cause light attenuation and further reduce image quality, and this impact is especially obvious in a hazy environment. Reduction of image quality results in the loss of information, which renders an image recognition system unable to identify objects in the image. In order to eliminate the hazy effect on images and improve the visual quality, this paper presents an efficient method combining the fuzzy inference system and the neural network filter to solve image dehazing. During dehazing, the fuzzy inference system is adopted to estimate the variations in light attenuation, and the erosion of morphological operation and the neural network filter are used to eliminate the halation and achieve optimization in transmission map refinement. Finally, the brightest 1% of the atmospheric light is utilized to calculate the color vector of atmospheric light to eliminate color cast. Experimental results indicate that the proposed method is superior to other dehazing methods. [ABSTRACT FROM AUTHOR]

In this paper, we investigate the effectiveness of applying deep neural network hidden Markov models, or DNN-HMMs, for acoustic modeling in the context of educational applications. Specifically, we focus on spoken responses from non-native and child speech that tend to show great acoustic variability. We perform comprehensive experiments to compare the performance between traditional Gaussian mixture model (GMM)-HMMs and DNN-HMMs in three large language assessment datasets that contain various spoken tasks, classified broadly as constrained and open-ended tasks. Our experimental results suggest useful conclusions that can help guide the design of real-life educational applications. DNN-HMMs outperform conventional GMM-HMMs by a large margin for all spoken tasks commonly used in spoken assessment applications. In our experiments, DNN-HMMs trained using 25 h of data can outperform GMM-HMMs trained with 6.7–9 times data. Specifically regarding overall performance, when all available training data were used (175, 227, 169 h respectively), we achieved a relative word error rate decrease of 20.4% for adult English and 29.3% for child English, and a relative character error rate decrease of 14.3% for adult Chinese, when switching from GMMs to DNNs. In comparing between types of tasks, we notice that the more challenging open-ended tasks benefit significantly more than constrained item types by the use of DNN-HMMs. For open-ended tasks, having large amounts of training data is the key, as DNN-HMMs can take full advantage of the added training data and further push performance. In contrast, the performance of constrained spoken tasks saturates at around 25 h of training data. At the same time, constrained spoken tasks require only a few hours of data (1 or 5 h) to build well-performing acoustic models. This is an encouraging observation, that indicates the potential to build reliable spoken assessment applications based on constrained tasks, when few domain specific training data are available. [ABSTRACT FROM AUTHOR]

This paper proposes a fuzzy cerebellar model articulation controller using a strategy-adaptation-based bacterial foraging optimization (SABFO) algorithm to solve classification problems. A strategic approach to the chemotaxis step in the SABFO algorithm was adopted: in this approach, each virtual bacterium swims on different run-lengths, and bacterial diversity is increased. The simulation results indicated that the performance of the proposed method was more favorable than that of other methods. [ABSTRACT FROM AUTHOR]

This study presents a new evolutionary learning algorithm to optimize the parameters of the neural fuzzy classifier (NFC). This new evolutionary learning algorithm is based on a hybrid of bacterial foraging optimization and particle swarm optimization. It is thus called bacterial foraging particle swarm optimization (BFPSO). The proposed BFPSO method performs local search through the chemotactic movement operation of bacterial foraging whereas the global search over the entire search space is accomplished by a particle swarm operator. The NFC model uses functional link neural networks as the consequent part of the fuzzy rules. This study uses orthogonal polynomials and linearly independent functions in a functional expansion of the functional link neural networks. Finally, the proposed neural fuzzy classifier with bacterial foraging particle swarm optimization (NFC-BFPSO) is adopted in several classification applications. Experimental results have demonstrated that the proposed NFC-BFPSO method can outperform other methods. [ABSTRACT FROM AUTHOR]

The aim of this paper is to distinguish the vehicle detection and count the class number in each classification from the inputs. We proposed the use of Fuzzy Guided Scale Choice (FGSC)-based SSD deep neural network architecture for vehicle detection and class counting with parameter optimization. The 'FGSC' blocks are integrated into the convolutional layers of the model, which emphasize essential features while ignoring less important ones that are not significant for the operation. We created the passing detection lines and class counting windows and connected them with the proposed FGSC-SSD deep neural network model. The 'FGSC' blocks in the convolution layer emphasize essential features and find out unnecessary features by using the scale choice method at the training stage and eliminate that significant speedup of the model. In addition, FGSC blocks avoided many unusable parameters in the saturation interval and improved the performance efficiency. In addition, the Fuzzy Sigmoid Function (FSF) increases the activation interval through fuzzy logic. While performing operations, the FGSC-SSD model reduces the computational complexity of convolutional layers and their parameters. As a result, the model tested Frames Per Second (FPS) on edge artificial intelligence (AI) and reached a real-time processing speed of 38.4 and an accuracy rate of more than 94%. Therefore, this work might be considered an improvement to the traffic monitoring approach by using edge AI applications. [ABSTRACT FROM AUTHOR]

A manufacturer's fabric first undergoes an abrasion test and manual visual inspection to grade the fabric prior to shipment to ensure that there are no defects present. Manual visual classification consumes a considerable amount of human resources. Furthermore, extended use of the eyes during visual inspection often causes occupational injuries, resulting in a decrease in the efficiency of the entire operation. In order to overcome and avoid such situations, this study proposed the use of deep principal components analysis-based neural networks (DPCANNs) for fabric pilling identification. In the proposed DPCANN, the characteristics of the hairball were automatically captured using deep principal components analysis (DPCA), and the hairball class was identified using the neural network and the support vector machine (SVM). The experimental results showed that the proposed DPCANN has an average accuracy of 99.7% at the hairball level, which is in line with the needs of the industry. The results also confirmed that the proposed hairball classification method is superior to other methods. [ABSTRACT FROM AUTHOR]

This study used Xilinx Field Programmable Gate Arrays (FPGAs) to implement a functional neuro-fuzzy network (FNFN) for solving nonlinear control problems. A functional link neural network (FLNN) was used as the consequent part of the proposed FNFN model. This study adopted the linear independent functions and the orthogonal polynomials in a functional expansion of the FLNN. Thus, the design of the FNFN model could improve the control accuracy. The learning algorithm of the FNFN model was divided into structure learning and parameter learning. The entropy measurement was adopted in the structure learning to determine the generated new fuzzy rule, whereas the gradient descent method in the parameter learning was used to adjust the parameters of the membership functions and the weights of the FLNN. In order to obtain high speed operation and real-time application, a very high speed integrated circuit hardware description language (VHDL) was used to design the FNFN controller and was implemented on FPGA. Finally, the experimental results demonstrated that the proposed hardware implementation of the FNFN model confirmed the viability in the temperature control of a water bath and the backing control of a car. [ABSTRACT FROM AUTHOR]

• We propose a unified two-stage framework for automatic model compression. • We propose Dynamic BN to predict the performance of each compression policy. • We propose a solving algorithm to search for the compression ratio allocation. Deep Neural Networks have become ubiquitous in various domains. Meanwhile, the problems of massive storage and computation costs have hindered the deployment of these models to real-world applications. This paper proposes a novel and unified two-stage framework for automatic model compression. To determine the compression ratio of each layer, we improve the optimization from two aspects. First, to predict the performance of each compression policy, we propose Dynamic BN, which improves the correlation significantly with little computation overhead. Second, to search for the compression ratio allocation, we propose an efficient and hyperparameter-free solving algorithm based on the proposed Hessian matrix approximation and Knapsack problem reformulation. Moreover, comprehensive experiments and analyses are conducted on the CIFAR-100&ImageNet datasets and various network architectures to demonstrate its performance advantages over existing model compression methods under the quantization-only, pruning-only, and pruning-quantization settings. [ABSTRACT FROM AUTHOR]

Visualization from trained deep neural networks has drawn massive public attention in recent. One of the visualization approaches is to train images maximizing the activation of specific neurons. However, directly maximizing the activation would lead to unrecognizable images, which cannot provide any meaningful information. In this paper, we introduce a simple but effective technique to constrain the optimization route of the visualization. By adding two totally inverse transformations, image blurring and deblurring, to the optimization procedure, recognizable images can be created. Our algorithm is good at extracting the details in the images, which are usually filtered by previous methods in the visualizations. Extensive experiments on AlexNet, VGGNet and GoogLeNet illustrate that we can better understand the neural networks utilizing the knowledge obtained by the visualization. [ABSTRACT FROM AUTHOR]

Channel Pruning is one of the most widespread techniques used to compress deep neural networks while maintaining their performances. Currently, a typical pruning algorithm leverages neural architecture search to directly find networks with a configurable width, the key step of which is to identify representative subnet for various pruning ratios by training a supernet. However, current methods mainly follow a serial training strategy to optimize supernet, which is very time-consuming. In this work, we introduce PSE-Net, a novel parallel-subnets estimator for efficient channel pruning. Specifically, we propose a parallel-subnets training algorithm that simulate the forward–backward pass of multiple subnets by droping extraneous features on batch dimension, thus various subnets could be trained in one round. Our proposed algorithm facilitates the efficiency of supernet training and equips the network with the ability to interpolate the accuracy of unsampled subnets, enabling PSE-Net to effectively evaluate and rank the subnets. Over the trained supernet, we develop a prior-distributed-based sampling algorithm to boost the performance of classical evolutionary search. Such algorithm utilizes the prior information of supernet training phase to assist in the search of optimal subnets while tackling the challenge of discovering samples that satisfy resource constraints due to the long-tail distribution of network configuration. Extensive experiments demonstrate PSE-Net outperforms previous state-of-the-art channel pruning methods on the ImageNet dataset while retaining superior supernet training efficiency. For example, under 300M FLOPs constraint, our pruned MobileNetV2 achieves 75.2% Top-1 accuracy on ImageNet dataset, exceeding the original MobileNetV2 by 2.6 units while only cost 30%/16% times than BCNet/AutoAlim. • We propose PSE-Net, a novel parallel-subnets estimator for efficient channel pruning. • The prior-distributed-based sampling algorithm to facilitate subnets search. • 30%/16% total time for estimator training compared with BCNet/AutoSlim. • Exceeds the SOTAs on ImageNet and downstream tasks for a variety of FLOPs budgets. [ABSTRACT FROM AUTHOR]