Search

This paper proposes a new technique to achieve a person's depth from a single image by considering the target foreground and background scene variations in extreme weather conditions within 40 meters range. For this purpose series of images are captured on each person at successive intervals. The height, distance, foreground, and background features are extracted using an object detection deep learning framework. The obtained features are then subsequently trained by the Gradient Booster Regressor to predict the depth information. Furthermore, the algorithm is tested on various images and is validated with ground truth depth data. The findings presented in this paper attest to the reliability of the methodology used for depth estimation.

With the rapid growth of threats and diversity in the manner of attack, Internet of things (IoT) systems has major challenges in providing methods to detect security vulnerabilities and attacks. There have been increasing developments of many detection tools and methods using full-time series data during malware execution based on machine learning/deep learning. However, the effectiveness of existing works is tightly bound by the requirement to use full-time series data. On the other hand, an earlier detection would help propose better solutions to respond to the IoT Botnet. Therefore, it mitigating the damage from potential attacks. In this paper, going beyond the full-time series data-based methods, we propose a collaborative machine learning model to effectively automate the early detection of IoT Botnet based on many features. The proposed model is 99.37% accurate on a dataset of 5023 IoT botnet and 3888 benign samples.

In this paper, we develop a feature extraction model using the amalgamation of machine and deep learning techniques for Industrial Internet of Thing (IIoT). We train the model with the most effective feature set evaluated using machine learning algorithms and deep learning feature extraction methods. We test these features with deep learning based network models for validation. We consider error metrics and accuracy as the major factors for combining the machine learning and deep learning techniques. The error rates are analysed using mean square error which show low error rate for the subset than the model tested for full dataset. Further, mean square error, accuracy rate and false values are analysed to test the performance of the proposed model. The comparative analysis with IIoT dataset and existing methods show that our approach is 25% better than the others.

Among the various reasons behind vehicle accidents, drivers' aggressiveness and distractions play a significant role. Deep learning (DL) algorithms inside a car mobile edge (CME) have been used for driver monitoring and to perform automated decision-making processes. Training and retraining the DL models in resource-constrained CME devices come with several challenges, especially regarding computational and memory space costs. Moreover, training the DL models periodically on representative data nearest to CME without imposing communication overheads on the cloud improves the quality of service (QoS) parameters, such as memory demand, processing time, power consumption, and bandwidth. This paper investigates the deployment of a deep neural network (DNN) model on a cloud-fog-edge computing framework for aggressive driver behavior detection and monitoring. To reach this goal, our framework proposes utilizing effective systems and databases of sensor-based metrics and data, cost-effective wireless networks, cloud-and fog-edge computing technologies, and the Internet. Experimental results of the DNN model showed that the accuracy of detection is improved by 1.84% compared with the current related work without any pre-processing window on data points that come from bio-signal sensors. Moreover, the experimental results of the networking part prove the efficiency and effectiveness of the proposed framework.

Cardiovascular disease has become a significant cause of modern people's health problems. The mobile electrocardiogram can be a non-intrusive and convenient tool to provide an auxiliary diagnosis of cardiovascular diseases for the virtual heath community. In this paper, a new diagnostic framework for mobile electrocardiogram signal is proposed with two improved methods. They are (1) the sliding segmentation method, which enhances the model's generalization ability and (2) the large-scale convolution kernel of a one-dimensional neural network, that is designed for mobile electrocardiogram signal with more resistant to sparseness and noise issues. According our experiments, the results show that the proposed sliding large kernel algorithm has a better accuracy than previous algorithms.

In this paper, the adversarial haze attack problem is addressed using the dark channel prior (DCP) de-hazing method. The adversarial attack affects rank-1 accuracy, where searching a target image against each test image is a specific search problem. To resolve this kind of problem, a feature fusion model is proposed to fuse handcrafted features and a pre-trained network model to obtain robust and discriminative features. The proposed model learns global features using transfer learning architecture whereas local features are obtained using the conventional method. Three pre-trained CNN models (AlexNet, ResNet, and Inception-v3) are used for feature extraction via transfer learning. The experiments are performed on publicly available datasets, achieving 68.6% accuracy in rank-1 with VIPER dataset and 79.6% accuracy with CHUK03 dataset. The proposed model enhances rank-1 accuracy of person re-identification when comparing with other state-of-the-art methods.

Human-Computer Interaction is the secret to technological advancement in the area of logistics and supply chain. The key challenges are the degree of energy transferred to devices, like automated vehicles and robotic equipment, and lack of belief in intelligent decision-making, which may overrule the system in the event of misperceptions of automated decisions. This paper presents an efficient Logistics Management Framework Using Deep Learning (eLMF-DL) to implement the computer vision-assisted Human-Computer Interaction (HCI) in the logistic management sector. With a hybrid CNN-LSTM network, eLMF-DL implements a single-stage or one-step convergence optimum decision-support design model that intelligently combines production maximization and demand forecasting. The architecture with the integration of convolutional neural network and long short-term memory network models the machine dynamics and relationships in assorted diverse logistics services demand. To determine uncertainties through dynamic delivery and optimal decisions on allocating logistical service power, the eLMF-DL results in the highest performance.

As an essential subproblem in the face recognition field, occluded face recognition has received considerable attention in recent years. However, satisfactory recognition accuracy for occluded faces has not been achieved. Therefore, in order to solve this problem, this paper proposes a convolutional neural network that has multidimensional serial feature extraction modules for occluded faces and uses the deep learning method to improve the recognition rate. In order to improve the expression ability of the network, the method extracts features from two dimensions — space and channel — and built a multidimensional serial feature extraction module. By "multiscale" and "dependence" processing features, the generalization ability was significantly improved. Through the experimental test, the recognition accuracy of the multidimensional feature network (MFNet) reached 90.35%. This accuracy is 17.89% (72.36%) better than that of the traditional algorithm. Meanwhile, compared with other convolutional neural networks, this network has different degrees of improvement, which are 0.68% (ArcFace), 2.14% (Visual Geometry Group Face), and 4.83% (DeepID3).

Recently, deep learning using Convolutional Neural Networks has played an essential role in many fields. Traditional cryptography, such as the technologies of the garbled circuit and the homomorphic encryption, may provide both parties with a private and secure computation in the neural networks as well as a secure inference scheme. However, it suffers heavy computation in practical designs especially for the training of a CNN model. Hence, the scalability of the model is restricted by these components. In this paper, we propose a novel deep learning model and a secure inferencing scheme in an application of a neural network. We utilize the inherent properties of a convolutional neural network to design a secure mechanism without using any complicated cryptography component. The security analysis shows our proposed scheme is secure, and the experimental results demonstrate that our method is very efficient and suitable for practical applications.

In this paper, we present an end-to-end approach method using deep learning for grasp detection. Our method is a real-time processing method for discrete depth image sampling and the problems of long calculation times and difficulty in registration caused by object modelling and global searching in traditional methods. The method uses depthwise convolution and pointwise convolution to model the relations among the channels and directly parameterizes a grasp quality value for every pixel. Our method calculates a rectangular grasping box to generate a grasping pose for an input image. For the experimental evaluation on the Jacquard dataset, we compared the proposed method with other baseline methods, and the accuracy of the proposed method was improved by 5% to 7% that shows our method can effectively predict grasp points on novel class objects.