L-SSD: lightweight SSD target detection based on depth-separable convolution

Journal of Real-Time Image Processing, Journal Year: 2024, Volume and Issue: 21(2)

Published: Feb. 16, 2024

Language: Английский

---

A novel robust adaptive neuro-sliding mode steering controller for autonomous ground vehicles

Robotics and Autonomous Systems, Journal Year: 2023, Volume and Issue: 170, P. 104557 - 104557

Published: Oct. 10, 2023

Language: Английский

---

An efficient adaptive feature selection with deep learning model-based paddy plant leaf disease classification

Multimedia Tools and Applications, Journal Year: 2023, Volume and Issue: 83(8), P. 22639 - 22661

Published: Aug. 7, 2023

Language: Английский

---

Video surveillance-based multi-task learning with swin transformer for earthwork activity classification

Engineering Applications of Artificial Intelligence, Journal Year: 2023, Volume and Issue: 131, P. 107814 - 107814

Published: Dec. 31, 2023

Language: Английский

---

Digital Twin technology for multimodal-based smart mobility using hybrid Co-ABC optimization based deep CNN

Cluster Computing, Journal Year: 2025, Volume and Issue: 28(3)

Published: Jan. 21, 2025

Language: Английский

---

Comprehensive review of single image defogging techniques: enhancement, prior, and learning based approaches

Artificial Intelligence Review, Journal Year: 2025, Volume and Issue: 58(4)

Published: Jan. 30, 2025

Language: Английский

---

Efficient Multi-Source Self-Attention Data Fusion for FDIA Detection in Smart Grid

Symmetry, Journal Year: 2023, Volume and Issue: 15(5), P. 1019 - 1019

Published: May 4, 2023

As a new cyber-attack method in power cyber physical systems, false-data-injection attacks (FDIAs) mainly disturb the operating state of systems by tampering with measurement data sensors, thereby avoiding bad-data detection grid and threatening security systems. However, existing FDIA methods usually only focus on feature extraction between false normal data, ignoring correlation that easily produces diverse redundancy, resulting significant difficulty detecting attacks. To address above problem, we propose multi-source self-attention fusion model for designing an efficient method. The proposed fusing firstly employs temporal alignment technique to integrate collected sensing identical time dimension. Subsequently, symmetric hybrid deep network is built symmetrically combining long short-term memory (LSTM) convolution neural (CNN), which can effectively extract features different data. Furthermore, design module further eliminate redundancy aggregate differences attack-data normal-data features. Finally, extracted their weights are integrated implement attack using single operation. Extensive simulations performed over IEEE14 node test IEEE118 systems; experimental results demonstrate our achieve better effects presents superior performance compared state-of-the-art.

Language: Английский

---

ENHANCING SURFACE DETECTION : A COMPREHENSIVE ANALYSIS OF VARIOUS YOLO MODELS

Heliyon, Journal Year: 2025, Volume and Issue: 11(3), P. e42433 - e42433

Published: Feb. 1, 2025

Material defects can significantly affect strength, durability and overall quality. Complex backgrounds variations in steel surface images often hinder productivity quality industrial environments. Accurate defect detection becomes difficult due to small target size unclear features. However, implementing accurate automated object algorithms mitigates these challenges, allowing errors or be detected before processing. Version 5 of You Only Look Once (YOLO), a precisely optimized learning model, has undergone extensive testing on strip datasets, providing effective solutions for recognition industry This study presents an improved YOLOv5 exploiting the efficient channel attention (ECA) coordinated (CoordAtt) mechanisms. Our results show notable improvements, with ECA hybrid mechanism achieving 2-4 times faster inference while maintaining high accuracy. Additionally, CoordAtt integration minimizes parameter count by 25% gives higher accuracy one datasets. Comparative analysis YOLOv6, YOLOv7, YOLOv8 demonstrates superior enhanced model NEU-DET GC10-DET benchmark highlighting its effectiveness detecting timely actual defects.

Language: Английский

---

Effective lane detection on complex roads with convolutional attention mechanism in autonomous vehicles

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: Aug. 19, 2024

Autonomous Vehicles (AV's) have achieved more popularity in vehicular technology recent years. For the development of secure and safe driving, these AV's help to reduce uncertainties such as crashes, heavy traffic, pedestrian behaviours, random objects, lane detection, different types roads their surrounding environments. In AV's, Lane Detection is one most important aspects which helps holding guidance departure warning. From Literature, it observed that existing deep learning models perform better on well maintained favourable weather conditions. However, performance extreme conditions curvy need focus. The proposed work focuses presenting an accurate detection approach poor roads, particularly those with curves, broken lanes, or no markings Convolutional Attention Mechanism (LD-CAM) model achieve this outcome. method comprises encoder, enhanced convolution block attention module (E-CBAM), a decoder. encoder unit extracts input image features, while E-CBAM quality feature maps images extracted from decoder provides output without loss any information original image. carried out using distinct data three datasets called Tusimple for condition images, Curve Lanes curve lanes Cracks Potholes damaged road images. trained showcased improved attaining Accuracy 97.90%, Precision 98.92%, F1-Score IoU 98.50% Dice Co-efficient 98.80% both structured defective

Language: Английский

---

A novel method for indian vehicle registration number plate detection and recognition using CNN

Journal of Intelligent & Fuzzy Systems, Journal Year: 2024, Volume and Issue: 46(4), P. 8563 - 8585

Published: Feb. 23, 2024

Automated reading of license plate and its detection is a crucial component the competent transportation system. Toll payment parking management e-payment systems may benefit from this software’s features. License identification algorithms abound, each has own set strengths weaknesses. Computer vision advanced rapidly in terms new breakthroughs techniques thanks to emergence proliferation deep learning principles across several branches AI. The practice automating monitoring process traffic management, police surveillance become much more effective development Automatic Plate Recognition (ALPR). Even though recognition (LPR) technology that extensively utilized been developed, there still significant amount work be done before it can achieve full potential. In last years, have substantial advancements both scientific community’s methodology level efficiency. era learning, numerous developments established for LPR, purpose research review examine those approaches. light this, authors study suggest four-stage technique automated (ALPDR), which includes, image pre-processing, extraction, character segmentation, recognition. And first three phases are known as “extraction,” “pre-processing,” “segmentation,” these processes shown unique technique. fact an essential detection, Convolution Neural Network (CNN), MobileNet, Inception V3, ResNet 50 all put through their paces regard.

Language: Английский

---