Human Face Detection Techniques: A Comprehensive Review and Future Research Directions

Electronics, Journal Year: 2021, Volume and Issue: 10(19), P. 2354 - 2354

Published: Sept. 26, 2021

Face detection, which is an effortless task for humans, complex to perform on machines. The recent veer proliferation of computational resources paving the way frantic advancement face detection technology. Many astutely developed algorithms have been proposed detect faces. However, there little attention paid in making a comprehensive survey available algorithms. This paper aims at providing fourfold discussions First, we explore wide variety five steps, including history, working procedure, advantages, limitations, and use other fields alongside detection. Secondly, include comparative evaluation among different each single method. Thirdly, provide detailed comparisons epitomized all-inclusive outlook. Lastly, conclude this study with several promising research directions pursue. Earlier papers are limited just technical details popularly used In our study, however, cover explanations various sub-branches neural network. We present under sub-branches. strengths limitations these novel literature that includes their besides

Language: Английский

---

A Collaborative Computation and Offloading for Compute-Intensive and Latency-Sensitive Dependency-Aware Tasks in Dew-Enabled Vehicular Fog Computing: A Federated Deep Q-Learning Approach

IEEE Transactions on Network and Service Management, Journal Year: 2023, Volume and Issue: 20(4), P. 4600 - 4614

Published: June 5, 2023

The demand for vehicular networks is prolifically emerging as it supports advancing in capabilities and qualities of vehicle services. However, this network cannot solely carry out latency-sensitive compute-intensive tasks, the slightest delay may cause any catastrophe. Therefore, fog computing can be a viable solution an integration to address aforementioned challenges. Moreover, complements Cloud reduces incurred latency ingress traffic by shifting resources edge network. This work investigated task offloading methods Vehicular Fog Computing (VFC) proposes Federated learning-supported Deep Q-Learning-based (FedDQL) technique optimal tasks collaborative paradigm. proposed method VFC performs computations, communications, offloading, resource utilization considering energy consumption. trade-offs between communication constraints were considered scenario. FedDQL scheme was validated dependent sets analyze efficacy method. Finally, results extensive simulations provide evidence that outperforms others with average improvement 49%, 34.3%, 29.2%, 16.2% 8.21%, respectively.

Language: Английский

---

An efficient fault tolerance scheme based enhanced firefly optimization for virtual machine placement in cloud computing

Concurrency and Computation Practice and Experience, Journal Year: 2023, Volume and Issue: 35(7)

Published: Jan. 17, 2023

Summary The virtual machine placement for the highly reliable cloud application is considered as one of challenging and critical issues. To tackle such an issue, this article proposes enhanced firefly algorithm based model. But migration time high to reduce placement, utilizes K‐means clustering algorithm. In addition, obtain optimal cluster adaptive particle swarm optimization with coyote employed. experimental results are conducted proposed approach using various measures transmission overhead, total execution time, packet size, parallel applications numbers, numbers. demonstrate that method offers improved performance scheme respect constraint factors. evaluation exposes less when compared other methods.

Language: Английский

---

RPRA: Reputation-based prioritization and resource allocation leveraging predictive analytics and vehicular fog computing

Ad Hoc Networks, Journal Year: 2024, Volume and Issue: 155, P. 103401 - 103401

Published: Jan. 11, 2024

Language: Английский

---

Evolution Characteristics and Causes—An Analysis of Urban Catering Cluster Spatial Structure

ISPRS International Journal of Geo-Information, Journal Year: 2023, Volume and Issue: 12(8), P. 302 - 302

Published: July 28, 2023

Studying the development characteristics of urban catering industry holds significant importance for understanding spatial patterns cities. In this manuscript, according to distribution points and based on point interest (POI) data 106 cities in China 2016 2022, we propose Natural Nearest Neighbor Single Branch Model (NNSBM) identify by adaptive clustering, which improves efficiency identifying clusters. Subsequently, a structure division model is constructed classify clusters into 3 major categories 17 subcategories, evolution pattern analyzed. addition, population density raster data, bivariate autocorrelation employed analyze complex relationship between density, revealing distinctive cluster evolution. The results showed that (1) initial stage formation, activities tend gather first specific area city, giving rise main cluster. However, as progresses, phenomenon “central fading” occurs within (2) overall trend most an toward low primacy–high concentration (Lp-Hc), at different stages capacity exhibited (3) influence was staged, with varying impact types structures.

Language: Английский

---

Optimizing Task Offloading for Collaborative Unmanned Aerial Vehicles (UAVs) in Fog–Cloud Computing Environments

IEEE Access, Journal Year: 2024, Volume and Issue: 12, P. 74698 - 74710

Published: Jan. 1, 2024

Unmanned Aerial Vehicles (UAVs) are used in various applications, including crowd management, crime prevention, accident detection, and rescue operations. However, since UAVs perform their tasks independently, some UAV applications dynamic geographically distributed, which may require extensive real-time processing capabilities. Thus, data locally can be challenging due to limited computing To overcome such limitations, fog cloud facilitate application development by providing additional resource capacities when needed. Despite this, designing sophisticated efficient task offloading strategies that collaborate with technologies considering service latency energy consumption, is rarely addressed the literature. Therefore, a collaborative strategy for presented this work, leveraging advantages This approach aims minimize UAVs' as well provide required resources services real time. In addition, decisions formulated using Mixed-Integer Linear Programming (MILP) model reduce consumption of entire UAV-fog-cloud system optimizing allocation computation communication requested each UAV. The simulation results demonstrate proposed significantly 15.38%, 35.29%, 59.26%, decrease overall (including networking) 3.3%, 7.37%, 12% compared alternative standalone (namely UAV, fog, cloud).

Language: Английский

---

ESIA: An Efficient and Stable Identity Authentication for Internet of Vehicles

IEEE Transactions on Vehicular Technology, Journal Year: 2023, Volume and Issue: 73(4), P. 5602 - 5615

Published: Nov. 6, 2023

Decentralized, tamper-proof blockchain is regarded as a solution to challenging authentication issue in the Internet of Vehicles (IoVs). However, consensus time and communication overhead increase significantly number vehicles connected blockchain. To address this issue, vehicular fog computing has been introduced improve efficiency. existing studies ignore several key factors such system, which can impact overhead. Meanwhile, there no comprehensive study on stability composition. The vehicle movement will lead dynamic changes fog. If composition unstable, formed by system be affect With above considerations, we propose an efficient stable identity ( ESIA ) empowered hierarchical computing. By grouping efficiently, low complexity achieves high stability. Moreover, enhance security blockchain, process from bottom layer up (bottom-up), call xmlns:xlink="http://www.w3.org/1999/xlink">B2UHChain . Through theoretical analysis simulation verification, our scheme design goals efficiency while improving IoV scalability power 1.5 (^1.5) under similar single-layer In addition, less computation overhead, lower latency, higher throughput than other baseline schemes.

Language: Английский

---

Green Demand Aware Fog Computing: A Prediction-Based Dynamic Resource Provisioning Approach

Electronics, Journal Year: 2022, Volume and Issue: 11(4), P. 608 - 608

Published: Feb. 16, 2022

Fog computing could potentially cause the next paradigm shift by extending cloud services to edge of network, bringing resources closer end-user. With its close proximity end-users and distributed nature, fog can significantly reduce latency. appearance more latency-stringent applications, in near future, we will witness an unprecedented amount demand for computing. Undoubtedly, this lead increase energy footprint network access segments. To consumption without compromising performance, paper propose Green-Demand-Aware Computing (GDAFC) solution. Our solution uses a prediction technique identify working nodes (nodes serve when request arrives), standby take over computational capacity is no longer sufficient), idle infrastructure. Additionally, it assigns appropriate sleep interval nodes, taking into account delay requirement applications. Results obtained based on mathematical formulation show that our save up 65% deteriorating performance.

Language: Английский

---

The Study of Mathematical Models and Algorithms for Face Recognition in Images Using Python in Proctoring System

Computation, Journal Year: 2022, Volume and Issue: 10(8), P. 136 - 136

Published: Aug. 9, 2022

The article analyzes the possibility and rationality of using proctoring technology in remote monitoring progress university students as a tool for identifying student. Proctoring includes face recognition technology. Face belongs to field artificial intelligence biometric recognition. It is very successful application image analysis understanding. To implement task determining person’s video stream, Python programming language was used with OpenCV code. Mathematical models are also described. These mathematical processed during data generation, classification. We considered methods that allow processes have presented algorithms solving computer vision problems. placed 400 photographs 40 on base. were taken at different angles lighting conditions; there interferences such presence beard, mustache, glasses, hats, etc. When analyzing certain cases errors, it can be concluded accuracy decreases primarily due images noise poor quality.

Language: Английский

---

A deep learning-based approach to facilitate the as-built state recognition of indoor construction works

Construction Innovation, Journal Year: 2022, Volume and Issue: 24(4), P. 933 - 949

Published: Dec. 19, 2022

Purpose Recognising the as-built state of construction elements is crucial for progress monitoring. Construction scholars have used computer vision-based algorithms to automate this process. Robust object recognition from indoor site images has been inhibited by technical challenges related objects, lighting conditions and camera positioning. Compared with traditional machine learning algorithms, one-stage detector deep (DL) can prioritise inference speed, enable real-time accurate detection classification. This study aims present a DL-based approach facilitate works. Design/methodology/approach The was built upon YOLO version 4 (YOLOv4) algorithm using transfer few hyperparameters customised trained in Google Colab virtual machine. process framing, insulation drywall installation partitions selected as scenario. For training, were captured two sites publicly available online images. Findings DL model reported best-trained weight mean average precision 92% an loss 0.83. previous studies, automation level high due use fixed time-lapse cameras data collection zero manual intervention pre-processing enhance visual quality Originality/value extends application models recognising works providing training Presenting workflow on platform reducing computational complexities associated also materialised.

Language: Английский

---