YOLO-FA: Type-1 fuzzy attention based YOLO detector for vehicle detection

Expert Systems with Applications, Год журнала: 2023, Номер 237, С. 121209 - 121209

Опубликована: Авг. 19, 2023

Язык: Английский

---

Hybrid Privacy Preserving Federated Learning Against Irregular Users in Next-Generation Internet of Things

Journal of Systems Architecture, Год журнала: 2024, Номер 148, С. 103088 - 103088

Опубликована: Фев. 17, 2024

Язык: Английский

---

Silent no more: a comprehensive review of artificial intelligence, deep learning, and machine learning in facilitating deaf and mute communication

Artificial Intelligence Review, Год журнала: 2024, Номер 57(7)

Опубликована: Июнь 26, 2024

Abstract People who often communicate via sign language are essential to our society and significantly contribute. They struggle with communication mostly because other people, do not understand language, cannot interact them. It is necessary develop a dependable system for automatic recognition. This paper aims provide comprehensive review of the advancements in artificial intelligence (AI), deep learning (DL), machine (ML) technologies that have been used facilitate individuals deaf mute (D–M). study explores various applications these technologies, including interpretation, speech recognition, text-to-speech synthesis. By examining current state research development AI, ML, DL D–M field, survey sheds light on potential challenges faced utilizing learning, ML bridge gap community. The findings this will contribute greater understanding impact improving access D–M, thereby aiding more inclusive accessible solutions.

Язык: Английский

---

Multiobjective Energy Consumption Optimization of a Flying–Walking Power Transmission Line Inspection Robot during Flight Missions Using Improved NSGA-II

Applied Sciences, Год журнала: 2024, Номер 14(4), С. 1637 - 1637

Опубликована: Фев. 18, 2024

In order to improve the flight efficiency of a flying–walking power transmission line inspection robot (FPTLIR) during missions, an accurate energy consumption model is constructed, and multiobjective optimization approach using improved NSGA-II proposed address high long execution time. The derived from FPTLIR kinematics motor dynamics, with key parameters validated test platform. A that considers many constraints related offering comprehensive analysis algorithm by integrating Cauchy variation operator simulated annealing algorithm, which used construct approach. Simulation experimental results demonstrate accurately predicts across different paths conditions average relative error ranging 0.76% 3.24%. After optimization, savings 5.33% 5.01% are achieved for on-line off-line respectively, while maintaining shortest time at given level. significantly improves system, providing reference analyzing optimizing robots.

Язык: Английский

---

Semi-supervised correntropy-based non-negative matrix factorization with hypergraph regularization

International Journal of Machine Learning and Cybernetics, Год журнала: 2025, Номер unknown

Опубликована: Янв. 17, 2025

Язык: Английский

---

Multi-objective optimization framework for generative design of horseshoe-shaped pipe arrangement in pre-stressed underground bundles

Tunnelling and Underground Space Technology, Год журнала: 2025, Номер 158, С. 106437 - 106437

Опубликована: Фев. 1, 2025

Язык: Английский

---

Exploring privacy measurement in federated learning

The Journal of Supercomputing, Год журнала: 2023, Номер 80(8), С. 10511 - 10551

Опубликована: Дек. 27, 2023

Язык: Английский

---

Study on Utilizing Mask R-CNN for Phenotypic Estimation of Lettuce’s Growth Status and Optimal Harvest Timing

Agronomy, Год журнала: 2024, Номер 14(6), С. 1271 - 1271

Опубликована: Июнь 12, 2024

Lettuce is an annual plant of the family Asteraceae. It most often grown as a leaf vegetable, but sometimes for its stem and seeds, growth status quality are evaluated based on morphological phenotypic traits. However, traditional measurement methods labor-intensive time-consuming due to manual measurements may result in less accuracy. In this study, we proposed new method utilizing RGB images Mask R-Convolutional Neural Network (CNN) estimating lettuce critical Leveraging publicly available datasets, employed improved R-CNN model perform analysis images. This allowed us estimate five traits simultaneously, which include fresh weight, dry height, canopy diameter, area. The enhanced involved two key aspects: (1) replacing backbone network from ResNet RepVGG enhance computational efficiency performance; (2) adding branches constructing multi-task regression achieve end-to-end estimation Experimental results demonstrated that present achieved high accuracy stable image segmentation, detection, tasks, with APs detection segmentation being 0.8684 0.8803, respectively. Additionally, R2 values 0.96, 0.9596, 0.9329, 0.9136, 0.9592, corresponding mean absolute percentage errors (MAPEs) 0.1072, 0.1522, 0.0757, 0.0548, 0.0899, study presents novel technical advancement digital knowledge evaluation quality, could lay foundation artificial intelligence expiation vegetable production.

Язык: Английский

---

A BERT-based recommender system for secure blockchain-based cyber physical drug supply chain management

Cluster Computing, Год журнала: 2023, Номер 26(6), С. 3389 - 3403

Опубликована: Июнь 23, 2023

Язык: Английский

---

Multi-Objective Optimization for Controlling the Dynamics of the Diabetic Population

Mathematics, Год журнала: 2023, Номер 11(13), С. 2957 - 2957

Опубликована: Июль 2, 2023

To limit the adverse effects of diabetes, a personalized and long-term management strategy that includes appropriate medication, exercise diet has become paramount importance necessity. Compartment-based mathematical control models for diabetes usually result in objective functions whose terms are conflicting, preventing use single-objective-based obtaining strategies. Taking into account conflicting aspects when controlling diabetic population dynamics, this paper introduces multi-objective approach consisting four steps: (a) modeling problem dynamics using model, (b) discretizing model trapezoidal rule Euler–Cauchy method, (c) swarm-intelligence-based optimizers to solve (d) structuring set controls soft clustering methods, known their flexibility. In contrast single-objective approaches, experimental results show obtains controls, where associated with compartment diabetics without complications is totally different from complications. Moreover, these enable significant reduction number complications, saves up 4% resources needed 18%

Язык: Английский

---