Novel Meta-Heuristic Algorithm for Feature Selection, Unconstrained Functions and Engineering Problems

The ever-growing available visual data (i.e., uploaded videos and pictures by internet users) has attracted the research community's attention in the computer vision field. Therefore, finding efficient solutions to extract knowledge from these sources is imperative. Recently, the BlazePose system has been released for skeleton extraction from images oriented to mobile devices. With this skeleton graph representation in place, a Spatial-Temporal Graph Convolutional Network can be implemented to predict the action. We hypothesize that just by changing the skeleton input data for a different set of joints that offers more information about the action of interest, it is possible to increase the performance of the Spatial-Temporal Graph Convolutional Network for HAR tasks. Hence, in this study, we present the first implementation of the BlazePose skeleton topology upon this architecture for action recognition. Moreover, we propose the Enhanced-BlazePose topology that can achieve better results than its predecessor. Additionally, we propose different skeleton detection thresholds that can improve the accuracy performance even further. We reached a top-1 accuracy performance of 40.1% on the Kinetics dataset. For the NTU-RGB+D dataset, we achieved 87.59% and 92.1% accuracy for Cross-Subject and Cross-View evaluation criteria, respectively.

Section: Related Workmentioning

confidence: 99%

Section: Ntu-rgb+dmentioning

confidence: 99%

Using BlazePose on Spatial Temporal Graph Convolutional Networks for Action Recognition

Alsawadi¹,

El-kenawy²,

Rio³

2023

“…DTO plays a crucial role in enhancing the performance of MLP by suppressing or strengthening the weights of the neural network. MLP is a nonlinear machine learning technique that is particularly useful in the detection phase since it is a stable, robust classifier with high detection accuracy [38]. Throughout the lifespan of the MANET, the proposed approach will be performed at regular intervals.…”

Section: Proposed Approachmentioning

confidence: 99%

Dipper Throated Optimization for Detecting Black-Hole Attacks in MANETs

Alkanhel¹,

El-kenawy²,

Abdelhamid³

et al. 2023

Self Cite

In terms of security and privacy, mobile ad-hoc network (MANET) continues to be in demand for additional debate and development. As more MANET applications become data-oriented, implementing a secure and reliable data transfer protocol becomes a major concern in the architecture. However, MANET's lack of infrastructure, unpredictable topology, and restricted resources, as well as the lack of a previously permitted trust relationship among connected nodes, contribute to the attack detection burden. A novel detection approach is presented in this paper to classify passive and active black-hole attacks. The proposed approach is based on the dipper throated optimization (DTO) algorithm, which presents a plausible path out of multiple paths for statistics transmission to boost MANETs' quality of service. A group of selected packet features will then be weighed by the DTO-based multi-layer perceptron (DTO-MLP), and these features are collected from nodes using the Low Energy Adaptive Clustering Hierarchical (LEACH) clustering technique. MLP is a powerful classifier and the DTO weight optimization method has a significant impact on improving the classification process by strengthening the weights of key features while suppressing the weights of minor features. This hybrid method is primarily designed to combat active black-hole assaults. Using the LEACH clustering phase, however, can also detect passive black-hole attacks. The effect of mobility variation on detection error and routing overhead is explored and evaluated using the suggested approach. For diverse mobility situations, the results demonstrate 1906 CMC, 2023, vol.74, no.1 up to 97% detection accuracy and faster execution time. Furthermore, the suggested approach uses an adjustable threshold value to make a correct conclusion regarding whether a node is malicious or benign.

“…The dataset contains a wide range of characteristics, both in size and unit. Data integrity may be maintained throughout the time when scaled [33][34][35][36][37][38][39][40]. As part of data preparation, the amount of samples in the dataset is balanced such that each class has equal numbers of samples [41][42][43][44][45][46][47].…”

Section: Dataset Preprocessingmentioning

confidence: 99%

Voting Classifier and Metaheuristic Optimization for Network Intrusion燚etection

Khafaga¹,

Karim²,

Abdelhamid³

et al. 2023

Managing physical objects in the network's periphery is made possible by the Internet of Things (IoT), revolutionizing human life. Open attacks and unauthorized access are possible with these IoT devices, which exchange data to enable remote access. These attacks are often detected using intrusion detection methodologies, although these systems' effectiveness and accuracy are subpar. This paper proposes a new voting classifier composed of an ensemble of machine learning models trained and optimized using metaheuristic optimization. The employed metaheuristic optimizer is a new version of the whale optimization algorithm (WOA), which is guided by the dipper throated optimizer (DTO) to improve the exploration process of the traditional WOA optimizer. The proposed voting classifier categorizes the network intrusions robustly and efficiently. To assess the proposed approach, a dataset created from IoT devices is employed to record the efficiency of the proposed algorithm for binary attack categorization. The dataset records are balanced using the locality-sensitive hashing (LSH) and Synthetic Minority Oversampling Technique (SMOTE). The evaluation of the achieved results is performed in terms of statistical analysis and visual plots to prove the proposed approach's effectiveness, stability, and significance. The achieved results confirmed the superiority of the proposed algorithm for the task of network intrusion detection.