Search

Sentiment analysis is inevitable in the current era. The Internet is growing day-byday. Now-a-days everything is online. We can shop, buy, and sell online. People can give feedback / opinions on the internet. Customers can compare among various products by analyzing the product reviews. As more and more people from different age groups and languages are becoming new internet users, we need it in regional languages. Till date most of the work related to sentiment analysis has been done in the English language. But when it comes to Indian languages, not much research has been done except for a few languages. This paper mainly focuses on performing sentiment analysis in one of the Indian languages i.e. Hindi. Lately, because of the accessibility of voluminous information on the web for Indian dialects, it has become a significant errand to break down this information to recover valuable data. In light of the development of Indian language content, it is helpful to use this blast of information with the end goal of conclusion investigation. This examination portrays a methodical survey in the field of opinion investigation all in all and Indian dialects explicitly. The current status of Indian dialects in estimation examination is grouped by the Indian language families. The periodical development of Indian dialects in the field of supposition examination, wellsprings of chose distributions based on their importance is additionally portrayed. Further, scientific categorization of Indian dialects in estimation examination dependent on strategies, areas, supposition levels and classes has been introduced. This examination work will help specialists in finding the accessible assets, for example, commented on datasets, pre-handling phonetic and lexical assets in Indian dialects for supposition investigation and will likewise uphold in choosing the most appropriate assumption examination strategy in a particular area alongside applicable future exploration bearings. In the event of asset helpless Indian dialects with morphological varieties, one experiences issues of performing estimation examination because of inaccessibility of explained assets, phonetic and lexical apparatuses. Along these lines, to give productive execution utilizing existing feeling examination procedures, the previously mentioned issues ought to be tended to successfully. [ABSTRACT FROM AUTHOR]

This paper proposes an automatic air-to-ground (A2G) channel model selection method based on machine learning (ML) using digital surface model (DSM) terrain data. In order to verify whether a communication network for a new non-terrestrial user service such as Urban Air Mobility (UAM) satisfies the required performance, it is necessary to perform a simulation reflecting the characteristics of the corresponding terrain environments as accurately as possible. For this simulation, A2G channel models corresponding to various terrain environments and a method of automatically classifying the terrain type of the simulation area must be provided. Many A2G channel models based on actual measurement results exist, but the practical automatic topography classification method still needs to be developed. This paper proposes the first practical automatic topography classification method using a two-step neural network-based classifier utilizing various geographic feature data as input. Since there is no open topography dataset to evaluate the accuracy of the proposed method, we built a new dataset for five topography classes that reflect the characteristics of Korea's topography, which is also a contribution of our study. The simulation results using the new data set show that the proposed ML-based method could increase the selection accuracy compared to the technique for direct classification by humans or the existing cross-correlation-based classification method. Since the proposed method utilizes the DSM data, open to the public, it can easily reflect the different terrain characteristics of each country. Therefore, the proposed method can be effectively used in the realistic performance evaluation of new non-terrestrial communication networks utilizing vast airspace such as UAM or 6G mobile communications. [ABSTRACT FROM AUTHOR]

With the development of the sharing economy, existing recommender services are changing from user-item recommendations to user-user recommendations. The most important consideration is that all users should have the best possible satisfaction. To achieve this outcome, the matching service adds information between users and items necessary for the existing recommender service and information between users, so higher-level data mining is required. To this end, this paper proposes a user-to-user matching service (UTU-MS) employing the prediction of mutual satisfaction based on learning. Users were divided into consumers and suppliers, and the properties considered for recommendations were set by filtering and weighting. Based on this process, we implemented a convolutional neural network (CNN)-deep neural network (DNN)-based model that can predict each supplier's satisfaction from the consumer perspective and each consumer's satisfaction from the supplier perspective. After deriving the final mutual satisfaction using the predicted satisfaction, a top recommendation list is recommended to all users. The proposed model was applied to match guests with hosts using Airbnb data, which is a representative sharing economy platform. The proposed model is meaningful in that it has been optimized for the sharing economy and recommendations that reflect user-specific priorities. [ABSTRACT FROM AUTHOR]

Phishing detection with high-performance accuracy and low computational complexity has always been a topic of great interest. New technologies have been developed to improve the phishing detection rate and reduce computational constraints in recent years. However, one solution is insufficient to address all problems caused by attackers in cyberspace. Therefore, the primary objective of this paper is to analyze the performance of various deep learning algorithms in detecting phishing activities. This analysis will help organizations or individuals select and adopt the proper solution according to their technological needs and specific applications' requirements to fight against phishing attacks. In this regard, an empirical study was conducted using four different deep learning algorithms, including deep neural network (DNN), convolutional neural network (CNN), Long Short-Term Memory (LSTM), and gated recurrent unit (GRU). To analyze the behaviors of these deep learning architectures, extensive experiments were carried out to examine the impact of parameter tuning on the performance accuracy of the deep learning models. In addition, various performance metrics were measured to evaluate the effectiveness and feasibility of DL models in detecting phishing activities. The results obtained from the experiments showed that no single DL algorithm achieved the best measures across all performance metrics. The empirical findings from this paper also manifest several issues and suggest future research directions related to deep learning in the phishing detection domain. [ABSTRACT FROM AUTHOR]

Summary: The omnipresence of drones in the civilian air space has led to their malicious usage raising high alert security issues. In this paper, a deep learning approach to detect and identify drones and to determine their flight modes from the remotely sensed radio frequency (RF) signatures is presented. This work intends to detect the presence of drones using two‐class classification, the presence along with identification of their make using four‐class classification. And this is further extended to the determination of their flight modes using ten‐class classification. It employs the proposed architectures of prominent deep learning classifiers, namely, autoencoder (AE), long short‐term memory (LSTM), convolutional neural network (CNN), and CNN‐LSTM hybrid model. To procure the relevant information from 227 RF signatures having 100 fragments each, the seven significant temporal statistical features, namely, maxima, minima, mean, variance, skewness, kurtosis, and root mean square, are extracted. In a two‐class classification scenario, all considered classifiers perform near to idle, whereas in a four‐class classification scenario, CNN performs best, followed by AE, CNN‐LSTM, and LSTM, respectively. Moreover, in a ten‐class classification scenario, AE far outperforms CNN, followed by LSTM and CNN‐LSTM, respectively. The best performance in terms of accuracy and classification time confirms the feasibility of the proposed AE classifier for the three considered drone operations. [ABSTRACT FROM AUTHOR]

In this paper, we propose a novel technique for the inspection of high-density polyethylene (HDPE) pipes using ultrasonic sensors, signal processing, and deep neural networks (DNNs). Specifically, we propose a technique that detects whether there is a diversion on a pipe or not. The proposed model transmits ultrasound signals through a pipe using a custom-designed array of piezoelectric transmitters and receivers. We propose to use the Zadoff–Chu sequence to modulate the input signals, then utilize its correlation properties to estimate the pipe channel response. The processed signal is then fed to a DNN that extracts the features and decides whether there is a diversion or not. The proposed technique demonstrates an average classification accuracy of 90.3 % (when one sensor is used) and 99.6 % (when two sensors are used) on 3 4 inch pipes. The technique can be readily generalized for pipes of different diameters and materials. [ABSTRACT FROM AUTHOR]

Vehicle fault detection and diagnosis (VFDD) along with predictive maintenance (PdM) are indispensable for early diagnosis in order to prevent severe accidents due to mechanical malfunction in urban environments. This paper proposes an early voiceprint driving fault identification system using machine learning algorithms for classification. Previous studies have examined driving fault identification, but less attention has focused on using voiceprint features to locate corresponding faults. This research uses 43 different common vehicle mechanical malfunction condition voiceprint signals to construct the dataset. These datasets were filtered by linear predictive coefficient (LPC) and wavelet transform(WT). After the original voiceprint fault sounds were filtered and obtained the main fault characteristics, the deep neural network (DNN), convolutional neural network (CNN), and long short-term memory (LSTM) architectures are used for identification. The experimental results show that the accuracy of the CNN algorithm is the best for the LPC dataset. In addition, for the wavelet dataset, DNN has the best performance in terms of identification performance and training time. After cross-comparison of experimental results, the wavelet algorithm combined with DNN can improve the identification accuracy by up to 16.57% compared with other deep learning algorithms and reduce the model training time by up to 21.5% compared with other algorithms. Realizing the cross-comparison of recognition results through various machine learning methods, it is possible for the vehicle to proactively remind the driver of the real-time potential hazard of vehicle machinery failure. [ABSTRACT FROM AUTHOR]

Lung cancer is considered to be one of the deadly diseases that threaten the survival of human beings. It is a challenging task to identify lung cancer in its early stage from the medical images because of the ambiguity in the lung regions. This paper proposes a new architecture to detect lung cancer obtained from the CT images. The proposed architecture has a three-stream network to extract the manual and automated features from the images. Among these three streams, automated feature extraction as well as the classification is done using residual deep neural network and custom deep neural network. Whereas the manual features are the handcrafted features obtained using high and low-frequency sub-bands in the frequency domain that are classified using a Support Vector Machine Classifier. This makes the architecture robust enough to capture all the important features required to classify lung cancer from the input image. Hence, there is no chance of missing feature information. Finally, all the obtained prediction scores are combined by weighted based fusion. The experimental results show 98.2% classification accuracy which is relatively higher in comparison to other existing methods. [ABSTRACT FROM AUTHOR]

Phishing detection with high-performance accuracy and low computational complexity has always been a topic of great interest. New technologies have been developed to improve the phishing detection rate and reduce computational constraints in recent years. However, one solution is insufficient to address all problems caused by attackers in cyberspace. Therefore, the primary objective of this paper is to analyze the performance of various deep learning algorithms in detecting phishing activities. This analysis will help organizations or individuals select and adopt the proper solution according to their technological needs and specific applications’ requirements to fight against phishing attacks. In this regard, an empirical study was conducted using four different deep learning algorithms, including deep neural network (DNN), convolutional neural network (CNN), Long Short-Term Memory (LSTM), and gated recurrent unit (GRU). To analyze the behaviors of these deep learning architectures, extensive experiments were carried out to examine the impact of parameter tuning on the performance accuracy of the deep learning models. In addition, various performance metrics were measured to evaluate the effectiveness and feasibility of DL models in detecting phishing activities. The results obtained from the experiments showed that no single DL algorithm achieved the best measures across all performance metrics. The empirical findings from this paper also manifest several issues and suggest future research directions related to deep learning in the phishing detection domain.

Detecting heart diseases has been a research interest for centuries. Many of these techniques are based on stethoscope, but only a few of these are digitally analyzed. In our paper, we propose a new method to detect heart diseases by analyzing heart sounds. Our goal is to help the medical doctor to identify whether a patient has heart disease or not. In general, doctors use acoustic stethoscope to detect abnormalities in the heart sound and predict abnormal conditions of the human heart. One major problem is that the frequency range and intensity of the heart sound is very low. Moreover, there are different types of heart sounds indicating different types of heart diseases. Hence, doctors are facing difficulties while detecting the cardiac sound and its abnormalities. Even the expert doctors may fail sometime to analyze heart sound properly. We developed and applied a novel data analysis to detect heart problems. Our method uses deep architecture features for analyzing heart diseases. We consider the heart sound as our raw data. This approach uses electronic stethoscope also known as e-stethoscope (that is, electronic stethoscope) to collect heart sounds and deep learning approach to identify that a heart has any disease or is healthy. If the heart has a disease, then it is desirable to identify the disease type. It is aimed to design a software known as a heartbeat audio classifier. This software should be able to differentiate normal heartbeats and heart murmurs which would assist the doctors to analyze a heart sound and detect a disease condition of the heart. Though our approach is not perfect, it shows that our approach leads to better results in comparison with others. [ABSTRACT FROM AUTHOR]

Deep Learning methods can deploy a fast, robust and lightweight model to solve the problem of 6-DOF camera relocalization in large-scale outdoor environments. However, two significant characteristics of captured images in a large-scale outdoor environment are moving objects, which should not include in the representation of an environment, and also motion blur which widely exists in the images captured with moving cameras. None of the existing approaches study and investigate these two problems in their method. This paper, for the first time, proposes a deep network architecture that is trained based on the knowledge achieved by combining deblurring and semantic segmentation modules and examines the effect of this combination on a challenging dataset. Results show approximately 20 and 50% improvement in camera position and orientation re-localization error respectively. [ABSTRACT FROM AUTHOR]

Usage of effective classification techniques on Magnetic Resonance Imaging (MRI) helps in the proper diagnosis of brain tumors. Previous studies have focused on the classification of normal (nontumorous) or abnormal (tumorous) brain MRIs using methods such as Support Vector Machine (SVM) and AlexNet. In this paper, deep learning architectures are used to classify brain MRI images into normal or abnormal. Gender and age are added as higher attributes for more accurate and meaningful classification. A deep learning Convolutional Neural Network (CNN)-based technique and a Deep Neural Network (DNN) are also proposed for effective classification. Other deep learning architectures such as LeNet, AlexNet, ResNet, and traditional approaches such as SVM are also implemented to analyze and compare the results. Age and gender biases are found to be more useful and play a key role in classification, and they can be considered essential factors in brain tumor analysis. It is also worth noting that, in most circumstances, the proposed technique outperforms both existing SVM and AlexNet. The overall accuracy obtained is 88% (LeNet Inspired Model) and 80% (CNN-DNN) compared to SVM (82%) and AlexNet (64%), with best accuracy of 100%, 92%, 92%, and 81%, respectively. [ABSTRACT FROM AUTHOR]

In recent years, research on convolutional neural networks (CNN) and recurrent neural networks (RNN) in deep learning has been actively conducted. In order to provide more personalized and advanced functions in smart home services, studies on deep learning applications are becoming more frequent, and deep learning is acknowledged as an efficient method for recognizing the voices and activities of users. In this context, this study aims to systematically review the smart home studies that apply CNN and RNN/LSTM as their main solution. Of the 632 studies retrieved from the Web of Science, Scopus, IEEE Explore, and PubMed databases, 43 studies were selected and analyzed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) methodology. In this paper, we examine which smart home applications CNN and RNN/LSTM are applied to and compare how they were implemented and evaluated. The selected studies dealt with a total of 15 application areas for smart homes, where activity recognition was covered the most. This study provides essential data for all researchers who want to apply deep learning for smart homes, identifies the main trends, and can help to guide design and evaluation decisions for particular smart home services. [ABSTRACT FROM AUTHOR]

Analog in-memory computing (AIMC) has been utilized in convolutional neural networks (CNNs) edge inference engines to solve the memory bottleneck problem and increase efficiency. However, AIMC analog-to-digital converters (ADCs) restricted resolution imposes quantization of output activations that can reduce the accuracy without meticulous optimization. A study conducted output quantization calibration and obtained configurations with which low-resolution ADCs did not affect the accuracy. The configurations were layer-specific. Therefore, a real-time quantization adjustment was required. AIMC output quantization is adjusted by controlling analog gain entangling it with analog parameters and nonlinear functions. AIMC dynamic output quantization control without interrupting its operation has been an unsettled problem until now. This paper introduces a technique for imposing output quantization configurations obtained from calibration processes on AIMC through circuit parameters setup. The technique permits on-the-fly quantization adjustments enabling layer-wise calibration that increases achievable network accuracies on AIMC platforms. As a case study, we deployed the method on the AIMC macro of an artificial intelligence (AI) inference engine SoC platform with a RISC-V processor and hybrid DIgital-ANAlog accelerators (DIANA). We related its controllable circuit parameters with the quantization configuration in a look-up table. This case study has noteworthy side benefits in identifying platform limitations due to nonlinearities and design imperfections. These limitations are investigated, and design advice that is transferable to future AIMC designs is provided to avoid imperfections such as mismatch, bias voltage drop, and interconnect delay. In addition, the study of output quantization from different levels of abstraction leads to design guidelines to facilitate dynamic quantization control during the application phase.

This paper introduces an off-grid DoA estimation via two-stage cascaded network which can resolve a mismatch between true direction-of-arrival (DoA) and discrete angular grid. In the first-stage network, the initial DoAs are estimated with a convolutional neural network (CNN), where initial DoAs are mapped on the discrete angular grid. To deal with the mismatch between initially estimated DoAs and true DoAs, the second-stage network estimates a tuning vector which represents the difference between true DoAs and nearest discrete angles. By using tuning vector, the final DoAs are estimated by moving initially estimated DoAs as much as the difference between true DoAs and nearest discrete angles. The limitation on estimation accuracy induced by the discrete angular grid can be resolved with the proposed two-stage network so that the estimation accuracy can be further enhanced. Simulation results show that adding the second-stage network after the first-stage network helps improve the estimation accuracy by resolving mismatch induced by the discretized grid. In the aspect of the implementation of machine learning, results also show that using CNN and using PReLU as the activation function is the best option for accurate estimation. [ABSTRACT FROM AUTHOR]

In this work, different deep learning approaches based on deep neural networks for person re-identification were analyzed. Both identification and re-identification of people are frequently required in various fields of human life. Some of the most common applications are in various security systems where it is necessary to identify and track a particular person. In the case of person identification, the identity of a particular person needs to be established. In the case of re-identification, the main task is to match the identity of a particular person across different, non-overlapping cameras or even with the same camera at different times. In this work, three different deep neural networks were used for the purpose of person re-identification. Two of them were user-defined, while one of them is a pre-trained neural network adapted to work with a specific dataset. Two neural networks used were Convolutional Neural Networks (CNN). For the defined experiment, it was used own dataset with 13 subjects in gait. [ABSTRACT FROM AUTHOR]

Lung cancer is considered to be one of the deadly diseases that threaten the survival of human beings. It is a challenging task to identify lung cancer in its early stage from the medical images because of the ambiguity in the lung regions. This paper proposes a new architecture to detect lung cancer obtained from the CT images. The proposed architecture has a three-stream network to extract the manual and automated features from the images. Among these three streams, automated feature extraction as well as the classification is done using residual deep neural network and custom deep neural network. Whereas the manual features are the handcrafted features obtained using high and low-frequency sub-bands in the frequency domain that are classified using a Support Vector Machine Classifier. This makes the architecture robust enough to capture all the important features required to classify lung cancer from the input image. Hence, there is no chance of missing feature information. Finally, all the obtained prediction scores are combined by weighted based fusion. The experimental results show 98.2% classification accuracy which is relatively higher in comparison to other existing methods.

Typhoon attacks on the Korean Peninsula have recently become more frequent, and the strength of these typhoons is also gradually increasing because of climate change. Typhoon attacks cause storm surges in coastal regions; therefore, forecasts that enable advanced preparation for these storm surges are important. Because storm surge forecasts require both accuracy and speed, this study uses an artificial neural network algorithm suitable for nonlinear modeling and rapid computation. A storm surge forecast model was created for five tidal stations on the Korea Strait (southern coast of the Korean Peninsula), and the accuracy of its forecasts was verified. The model consisted of a deep neural network and convolutional neural network that represent the two-dimensional spatial characteristics. Data from the Global Forecast System numerical weather model were used as input to represent the spatial characteristics. The verification of the forecast accuracy revealed an absolute relative error of ≤5% for the five tidal stations. Therefore, it appears that the proposed method can be used for forecasts for other locations in the Korea Strait. Furthermore, because accurate forecasts can be computed quickly, the method is expected to provide rapid information for use in the field to support advance preparation for storm surges. [ABSTRACT FROM AUTHOR]

• Filters are visualized by the activation maximization to explain functions of filters. • DNNs are compressed based on the visualization results. • The redundant filters are measured based on the color and texture similarities. • The repetitive and invalid filters can be pruned by optimization. This paper proposes a method to compress deep neural networks (DNNs) based on interpretability. For a trained DNN model, the activation maximization technique is first used to visualize every filter of the DNN model. Then, a single-layer filter pruning approach is introduced from what is learned by visualization. The entire DNN model is compressed layer by layer by using the single-layer filter pruning method in which the compression of the current layer is based on the compression of the preceding layers. Importantly, in addition to effective compression, the proposed method renders a better interpretation of the deep learning process. With a 60 % compression rate of the VGG-16, our method achieves 0.8429 Top-1 accuracy under CIFAR-10, with a slight accuracy drop of only 0.0322, and the storage space of the model can be compressed to 9.42 Mb. For a modern DNN model such as ResNet50, our visualization-based filter pruning method is significantly better than other pruning strategies in different convolutional layers under different compression rates and the larger ImageNet dataset. After pruning, the computation cost and storage requirement of the DNN can be significantly reduced, which means that complex DNN models can be easily implemented in small mobile devices, thus enabling the efficient use of DNNs in the Internet of Things technologies. [ABSTRACT FROM AUTHOR]

The human brain consists of billions of living organisms and is very difficult to decipher because of its complexity. Brain tumors can be deadly, significantly impacting the quality of life and changing everything for patients and their loved ones. In today’s world, brain tumors are a leading cause of death in both children and adults. A high death percentage is caused due to the invasive properties of tumors. But it is inspiring that the survival rate might increase if the diagnosis is performed at the early stage [9]. Accurate detection of the brain tumor at an early stage can prolong the chance of survival of an infected patient [4]. Magnetic Resonance Imaging (MRI) is the most popular imaging technique used today for detecting brain tumors. Deep Neural Network techniques plays an important role in detecting brain tumors. This manuscript offers a brief analysis of studies conducted by various authors in the field of BT categorization and diagnosis from MRI images using Deep Neural Network (DNN). This paper also suggests a method for classifying and identifying brain tumors based on MRI pictures and supporting text using DNN and DWT.

AC series arc is dangerous and can cause serious electric fire hazards and property damage. This article proposed a convolutional neural network -based arc detection model named ArcNet. The database of this research is collected from eight different types of loads according to IEC62606 standard. The two most common types of arcs, including arcs from a loose connection of cables and those caused by the failure of the insulation, are generated in testing and included in the database. Using the database of raw current, experimental results indicate ArcNet can achieve a maximum of 99.47% arc detection accuracy at 10 kHz sampling rate. The model is also implemented in Raspberry Pi 3B for classification accuracy. A tradeoff study between the arc detection accuracy and model runtime has been conducted. The proposed ArcNet obtained an average runtime of 31 ms/sample of 1 cycle at 10 kHz sampling rate, which proves the feasibility of practical hardware deployment for real-time processing. [ABSTRACT FROM AUTHOR]

In this paper, we propose a novel technique for the inspection of high-density polyethylene (HDPE) pipes using ultrasonic sensors, signal processing, and deep neural networks (DNNs). Specifically, we propose a technique that detects whether there is a diversion on a pipe or not. The proposed model transmits ultrasound signals through a pipe using a custom-designed array of piezoelectric transmitters and receivers. We propose to use the Zadoff–Chu sequence to modulate the input signals, then utilize its correlation properties to estimate the pipe channel response. The processed signal is then fed to a DNN that extracts the features and decides whether there is a diversion or not. The proposed technique demonstrates an average classification accuracy of 90.3% (when one sensor is used) and 99.6% (when two sensors are used) on 34 inch pipes. The technique can be readily generalized for pipes of different diameters and materials.

Vehicle fault detection and diagnosis (VFDD) along with predictive maintenance (PdM) are indispensable for early diagnosis in order to prevent severe accidents due to mechanical malfunction in urban environments. This paper proposes an early voiceprint driving fault identification system using machine learning algorithms for classification. Previous studies have examined driving fault identification, but less attention has focused on using voiceprint features to locate corresponding faults. This research uses 43 different common vehicle mechanical malfunction condition voiceprint signals to construct the dataset. These datasets were filtered by linear predictive coefficient (LPC) and wavelet transform(WT). After the original voiceprint fault sounds were filtered and obtained the main fault characteristics, the deep neural network (DNN), convolutional neural network (CNN), and long short-term memory (LSTM) architectures are used for identification. The experimental results show that the accuracy of the CNN algorithm is the best for the LPC dataset. In addition, for the wavelet dataset, DNN has the best performance in terms of identification performance and training time. After cross-comparison of experimental results, the wavelet algorithm combined with DNN can improve the identification accuracy by up to 16.57% compared with other deep learning algorithms and reduce the model training time by up to 21.5% compared with other algorithms. Realizing the cross-comparison of recognition results through various machine learning methods, it is possible for the vehicle to proactively remind the driver of the real-time potential hazard of vehicle machinery failure.

Usage of effective classification techniques on Magnetic Resonance Imaging (MRI) helps in the proper diagnosis of brain tumors. Previous studies have focused on the classification of normal (nontumorous) or abnormal (tumorous) brain MRIs using methods such as Support Vector Machine (SVM) and AlexNet. In this paper, deep learning architectures are used to classify brain MRI images into normal or abnormal. Gender and age are added as higher attributes for more accurate and meaningful classification. A deep learning Convolutional Neural Network (CNN)-based technique and a Deep Neural Network (DNN) are also proposed for effective classification. Other deep learning architectures such as LeNet, AlexNet, ResNet, and traditional approaches such as SVM are also implemented to analyze and compare the results. Age and gender biases are found to be more useful and play a key role in classification, and they can be considered essential factors in brain tumor analysis. It is also worth noting that, in most circumstances, the proposed technique outperforms both existing SVM and AlexNet. The overall accuracy obtained is 88% (LeNet Inspired Model) and 80% (CNN-DNN) compared to SVM (82%) and AlexNet (64%), with best accuracy of 100%, 92%, 92%, and 81%, respectively.

In recent years, research on convolutional neural networks (CNN) and recurrent neural networks (RNN) in deep learning has been actively conducted. In order to provide more personalized and advanced functions in smart home services, studies on deep learning applications are becoming more frequent, and deep learning is acknowledged as an efficient method for recognizing the voices and activities of users. In this context, this study aims to systematically review the smart home studies that apply CNN and RNN/LSTM as their main solution. Of the 632 studies retrieved from the Web of Science, Scopus, IEEE Explore, and PubMed databases, 43 studies were selected and analyzed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) methodology. In this paper, we examine which smart home applications CNN and RNN/LSTM are applied to and compare how they were implemented and evaluated. The selected studies dealt with a total of 15 application areas for smart homes, where activity recognition was covered the most. This study provides essential data for all researchers who want to apply deep learning for smart homes, identifies the main trends, and can help to guide design and evaluation decisions for particular smart home services.

An accurate head-related transfer function can improve the subjective auditory localization performance of a particular subject. This paper proposes a deep neural network model for reconstructing the head-related transfer function (HRTF) based on anthropometric parameters and the orientation of the sound source. The proposed model consists of three subnetworks, including a one-dimensional convolutional neural network (1D-CNN) to process anthropometric parameters as input features and another network that takes the sound source position as input to serve as a marker. Finally, the outputs of these two networks are merged together as the input to a third network to estimate the HRTF. An objective method and a subjective method are proposed to evaluate the performance of the proposed method. For the objective evaluation, the root mean square error (RMSE) between the estimated HRTF and the measured HRTF is calculated. The results show that the proposed method performs better than a database matching method and a deep-neural-network-based method. In addition, the results of a sound localization test performed for the subjective evaluation show that the proposed method can localize sound sources with higher accuracy than the KEMAR dummy head HRTF or the DNN-based method. The objective and subjective results all show that the personalized HRTFs obtained using the proposed method perform well in HRTF reconstruction. [ABSTRACT FROM AUTHOR]

Deep neural networks (DNNs) are increasing their presence in a wide range of applications, and their computationally intensive and memory-demanding nature poses challenges, especially for embedded systems. Pruning techniques turn DNN models into sparse by setting most weights to zero, offering optimization opportunities if specific support is included. We propose a novel pipelined architecture for DNNs that avoids all useless operations during the inference process. It has been implemented in a field-programmable gate array (FPGA), and the performance, energy efficiency, and area have been characterized. Exploiting sparsity yields remarkable speedups but also produces area overheads. We have evaluated this tradeoff in order to identify in which scenarios it is better to use that area to exploit sparsity, or to include more computational resources in a conventional DNN architecture. We have also explored different arithmetic bitwidths. Our sparse architecture is clearly superior on 32-bit arithmetic or highly sparse networks. However, on 8-bit arithmetic or networks with low sparsity it is more profitable to deploy a dense architecture with more arithmetic resources than including support for sparsity. We consider that FPGAs are the natural target for DNN sparse accelerators since they can be loaded at run-time with the best-fitting accelerator. [ABSTRACT FROM AUTHOR]

Under‐display camera is of great interest in the display industry potentially eliminating the display bezel and camera notch/hole in mobile devices. However, display panels cause complex signal modulation in the camera aperture which results in obscuration, attenuation and diffraction of the incident light. We propose a learning‐based image restoration approach to enable a camera to operate underneath the display without affecting the display contrast and color gamut. [ABSTRACT FROM AUTHOR]

This paper introduces an off-grid DoA estimation via two-stage cascaded network which can resolve a mismatch between true direction-of-arrival (DoA) and discrete angular grid. In the first-stage network, the initial DoAs are estimated with a convolutional neural network (CNN), where initial DoAs are mapped on the discrete angular grid. To deal with the mismatch between initially estimated DoAs and true DoAs, the second-stage network estimates a tuning vector which represents the difference between true DoAs and nearest discrete angles. By using tuning vector, the final DoAs are estimated by moving initially estimated DoAs as much as the difference between true DoAs and nearest discrete angles. The limitation on estimation accuracy induced by the discrete angular grid can be resolved with the proposed two-stage network so that the estimation accuracy can be further enhanced. Simulation results show that adding the second-stage network after the first-stage network helps improve the estimation accuracy by resolving mismatch induced by the discretized grid. In the aspect of the implementation of machine learning, results also show that using CNN and using PReLU as the activation function is the best option for accurate estimation.

Distance information of an obstacle is important for obstacle avoidance in many applications, and could be used to determine the potential risk of object collision. In this study, the detection of a moving fixed-wing unmanned aerial vehicle (UAV) with deep learning-based distance estimation to conduct a feasibility study of sense and avoid (SAA) and mid-air collision avoidance of UAVs is proposed by using a monocular camera to detect and track an incoming UAV. A quadrotor is regarded as an owned UAV, and it is able to estimate the distance of an incoming fixed-wing intruder. The adopted object detection method is based on the you only look once (YOLO) object detector. Deep neural network (DNN) and convolutional neural network (CNN) methods are applied to exam their performance in the distance estimation of moving objects. The feature extraction of fixed-wing UAVs is based on the VGG-16 model, and then its result is applied to the distance network to estimate the object distance. The proposed model is trained by using synthetic images from animation software and validated by using both synthetic and real flight videos. The results show that the proposed active vision-based scheme is able to detect and track a moving UAV with high detection accuracy and low distance errors. [ABSTRACT FROM AUTHOR]

Natural interaction in videoconferences can be made by correcting gaze/perspective, scale, and position by using cameras placed behind a partially transparent front emitting OLED panel and frame interpolation of deep neural networks. The diffraction artifacts from through‐screen imaging are removed by a deconvolution method. [ABSTRACT FROM AUTHOR]