Deep Learning-based Detection Model Research Articles

Heterogeneous Ensemble Approaches for Robust Face Mask Detection in Crowd Scenes

Face masks are one of the effective tools to slow the spread of disease and reduce medical overload by protecting people from infectious diseases including COVID-19. To prevent infection from respiratory droplets, it is imperative to wear a mask that covers the nose and mouth completely. However, it is difficult to make it mandatory for crowds to wear masks in public places where many people gather. For example, detecting incorrect mask-wearing in crowded scenes is a tedious and attention-grabbing task. Therefore, the success of deep learning in computer vision motivates automated monitoring systems. However, deep learning-based detection models are unstable if the domain task is changed and may have different strengths and weaknesses. Therefore, in this study, we propose a heterogeneous ensemble-based detection model for robust face mask detection in crowd scenes. First, independent detection models such as You look Only Ones (YOLO) v6, YOLO v7, and Faster R-CNN are employed for the model ensemble. Second, the prediction results obtained from the detection models are post-processed such as merging, non-maximum suppression, and weighted box fusion. The experimental results show that the classification performance of our proposed model has an F1 score of about 90.5% and that the improvement of the generalization ability due to the ensemble strategy contributed to the improvement of the classification performance.

A Covert Electricity-Theft Cyberattack Against Machine Learning-Based Detection Models

The advanced metering infrastructure in modern networked smart homes brings various advantages, such as multiple pricing and energy scheduling. However, smart homes are vulnerable to many cyberattacks, and the most striking one is energy theft. Researchers have developed many countermeasures, fostered by advanced machine learning (ML) techniques. Nevertheless, recent advances are not robust enough in practice, partially due to the vulnerabilities of employed ML algorithms. Given a group of smart homes, in this article, we present a covert electricity theft strategy through mimicking normal consumption patterns and compromising neighboring meters concurrently. Such attack is almost impossible to be detected by existing solutions as the manipulated data have little deviation against honest usage records. To address this threat, we initially identify and define two levels of consumption deviations: home-level and interpersonal-level, respectively. Then, we design a feature extraction scheme that can capture the correlation between attacks and honest customers. Finally, we develop a new deep learning-based detection model. Extensive experiments based on real-world datasets show that the presented attack could evade existing mainstream detectors but still gain high profits. In addition, the proposed countermeasure outperforms state-of-the-art detection methods.

Detection and tracking for the awareness of surroundings of a ship based on deep learning

Abstract To prevent maritime accidents, it is crucial to be aware of the surrounding environment near ships. The images recorded by a camera mounted on a ship could be used for the awareness of other ships surrounding it. In this study, ship awareness was performed using three procedures: detection, localization, and tracking. Initially, ship detection was performed using the deep learning-based detection model, YOLO (You Only Look Once) v3, based on the camera image. A virtual image dataset was constructed using Unity to overcome the difficulty of obtaining camera images onboard with various sizes of ships, and to improve the detection performance. This was followed by the localization procedure in which the position of the horizon on the image was calculated using the orientation information from the ship. Subsequently, the position of the detected ship in the spatial coordinate system was calculated using the horizon information. Following this, the position, course over ground, and speed over ground of the target ships were tracked in the time domain using the extended Kalman filter. A deep learning model that determines the heading of the ship in the image was proposed to utilize abundant information of cameras, and it was used to set the initial value of the Kalman filter. Finally, the proposed method for the awareness of ships was validated using an actual video captured from a camera installed on an actual ship with various encountering scenarios. The tracking results were compared with actual automatic identification system data obtained from other ships. As a result, the entire detection, localization, and tracking procedures showed good performance, and it was estimated that the proposed method for the awareness of the surroundings of a ship, based on camera images, could be used in the future.

Semi-supervised learning approach for malicious URL detection via adversarial learning1

Uniform Resource Location (URL) is the network unified resource location system that specifies the location and access method of resources on the Internet. At present, malicious URL has become one of the main means of network attack. How to detect malicious URL timely and accurately has become an engaging research topic. The recent proposed deep learning-based detection models can achieve high accuracy in simulations, but several problems are exposed when they are used in real applications. These models need a balanced labeled dataset for training, while collecting large numbers of the latest labeled URL samples is difficult due to the rapid generation of URL in the real application environment. In addition, in most randomly collected datasets, the number of benign URL samples and malicious URL samples is extremely unbalanced, as malicious URL samples are often rare. This paper proposes a semi-supervised learning malicious URL detection method based on generative adversarial network (GAN) to solve the above two problems. By utilizing the unlabeled URLs for model training in a semi-supervised way, the requirement of large numbers of labeled samples is weakened. And the imbalance problem can be relieved with the synthetic malicious URL generated by adversarial learning. Experimental results show that the proposed method outperforms the classic SVM and LSTM based methods. Specially, the proposed method can obtain high accuracy with insufficient labeled samples and unbalanced dataset. e.g., the proposed method can achieve 87.8% /91.9% detection accuracy when the number of labeled samples is reduced to 20% /40% of that of conventional methods.

A malicious threat detection model for cloud assisted internet of things (CoT) based industrial control system (ICS) networks using deep belief network

Internet of Things (IoT) devices are extensively used in modern industries combined with the conventional industrial control system (ICS) network through the industrial cloud to make the production data easily available to the corporate business management and easier control for highly profitable production systems. The different devices within the conventional ICS network originally manufactured to run on an isolated network and was not considered for the privacy and security of the control and production/architecture data being trafficked over the manufacturing plant to the corporate. Due to their extensive integration with the industrial cloud network over the internet, these ICS networks are exposed to a significant threat of malicious activities created by malicious software. Protecting ICS from such attacks requires continuous update of their database of anti-malware tools which requires efforts from manual experts on a regular basis. This limits real time protection of ICS.Earlier work by Huda et al. (2017) based on a semi-supervised approach performed well. However training process of the semi-supervised-approach (Huda et al., 2017) is complex procedure which requires a hybridization of feature selection, unsupervised clustering and supervised training techniques. Therefore, it could be time consuming for ICS network for real time protection. In this paper, we propose an adaptive threat detection model for industrial cloud of things (CoT) based on deep learning. Deep learning has been used in many domain of pattern recognition and a popular approach for its simple training procedure. Most importantly, deep learning can learn the hidden patterns of the domain in an unsupervised manner which can avoid the requirements of huge expensive labeled data. We used this particular characteristic of deep learning to design our detection model.Two different types of deep learning based detection models are proposed in this work. The first model uses a disjoint training and testing data for a deep belief network (DBN) and corresponding artificial neural network (ANN). In the second proposed detection model, DBN is trained using new unlabeled data to provide DBN with additional knowledge about the changes in the malicious attack patterns. Novelty of the proposed detection models is that the models are adaptive where training procedures is simpler than earlier work (Huda et al, 2017) and can adapt new malware behaviors from already available and cheap unlabeled data at the same time. This will avoid expensive manual labeling of new attacks and corresponding time complexity making it feasible for ICS networks. Performances of standard DBNs are sensitive to its configurations and values for the hyper-parameters including number of hidden nodes, learning rate and number epochs. Therefore proposed detection models find an optimal configuration by varying the structure of DBNs and other parameters. The proposed detection models are extensively tested on a real malware test bed. Experimental results show that the proposed approaches achieve higher accuracies than standard detection algorithms and obtain similar performances with earlier semi-supervised work (Huda et al., 2017) but provide a comparatively simplified training model.