Bioinspired Garra Rufa Optimization-Assisted Deep Learning Model for Object Classification on Pedestrian Walkways DOI Creative Commons
Eunmok Yang, K. Shankar, Sachin Kumar

et al.

Biomimetics, Journal Year: 2023, Volume and Issue: 8(7), P. 541 - 541

Published: Nov. 11, 2023

Object detection in pedestrian walkways is a crucial area of research that widely used to improve the safety pedestrians. It not only challenging but also tedious process manually examine labeling abnormal actions, owing its broad applications video surveillance systems and larger number videos captured. Thus, an automatic system identifies anomalies has become indispensable for computer vision (CV) researcher workers. The recent advancements deep learning (DL) algorithms have attracted wide attention CV processes such as object classification based on supervised requires labels. current study designs bioinspired Garra rufa optimization-assisted model (BGRODL-OC) technique walkways. objective BGRODL-OC recognize presence pedestrians objects video. To achieve this goal, primarily applies GhostNet feature extractors produce set vectors. In addition this, makes use GRO algorithm hyperparameter tuning process. Finally, performed via attention-based long short-term memory (ALSTM) network. A range experimental analysis was conducted validate superior performance technique. values established over other existing approaches.

Language: Английский

Towards walkable footpath detection for the visually impaired on Bangladeshi roads with smartphones using deep edge intelligence DOI Creative Commons
Md. Ishan Arefin Hossain, Jareen Anjom, Rashik Iram Chowdhury

et al.

Array, Journal Year: 2025, Volume and Issue: unknown, P. 100388 - 100388

Published: April 1, 2025

Language: Английский

Citations

0
Astute Assistance System for Blind and Visually Impaired People DOI

P. Usha Rani,

S Angel,

L Janani

et al.

Published: March 11, 2024

There are numerous hand-held obstacle detectors and ultrasonic guide devices available for individuals with visual impairments, all of which do not cause injury allow them to cross roads safely. To assist while crossing the roads, this study proposes a novel solution at traffic junctions by utilizing listener commands. The proposed model works help pedestrians activating control system, temporarily pausing signals, directing when cross.

Language: Английский

Citations

2
Intelligent Multi-Group Marine Predator Algorithm With Deep Learning Assisted Anomaly Detection in Pedestrian Walkways DOI Creative Commons

S. Rama Sree,

E. Laxmi Lydia, Mohammed Altaf Ahmed

et al.

IEEE Access, Journal Year: 2024, Volume and Issue: 12, P. 72662 - 72671

Published: Jan. 1, 2024

Anomaly Detection (AD) in Pedestrian Walkways (PWs) is critical to urban security and safety systems. It widely used detect abnormal or unusual behaviours, situations, events areas dedicated pedestrian traffic, namely crosswalks, sidewalks, bridges. The main objective improve efficiency, safety, the environment by identifying deviations monitoring activities from established norms. This kind of AD typically includes surveillance cameras, sensors, advanced software algorithms. Using machine learning (ML) computer vision (CV) approaches, this technique continuously monitors area potential threats irregularities. Deep Learning Assisted presents a novel very efficient method enhance environments. Therefore, study designs an Intelligent Multi-Group Marine Predator Algorithm with (MMPADL-AD) Walkways. MMPADL-AD system aims ensure PWs via process. incorporates NASNet feature extractor that proficiently extracts high-level features data, allowing deep understanding behaviours. Besides, applies convolutional long short-term memory (ConvLSTM), inheriting benefits neural networks) LSTM for Finally, MMPA has been hyperparameter tuning mechanism, which optimizes model's performance, assuring accuracy adaptability. Benchmark data accompanied extensive set experiments higher effectiveness approach. experimental values highlighted supremacy approach over other DL methods.

Language: Английский

Citations

0
Blind assistance system for appliance control and public transport safety using CNN, MobileNet V2 and Yolo V8 DOI

S. Srividhya,

V. Brindha

International Journal of Hybrid Intelligent Systems, Journal Year: 2024, Volume and Issue: 20(3), P. 243 - 258

Published: June 28, 2024

The purpose of this research is to address the challenges faced by visually impaired individuals, particularly in handling household appliances independently. With approximately 285 million individuals worldwide, technological solutions are crucial enhancing their accessibility and independence. This paper introduces a Smart Assistance System designed empower interact with real-time without assistance. In study, three Convolutional Neural Network (CNN) algorithms compared develop system. evaluation metrics include accuracy, precision, recall, F1 score, hamming loss on validation images. performance comparison reveals that custom architecture CNN, MobileNetv2, YOLO models achieve scores 0.43, 0.63, 0.24, respectively. To enhance object detection classification, suggests implementing bounding box buttons categorization using YOLOv8, which demonstrates superior 95% classification accuracy testing images home appliance buttons. They face similar difficult while public accessing property. Expanding upon proposed system’s capabilities, concept panic button activation bus environment tailored for blind individuals. system relies various factors such as number people onboard, heart rate monitoring, distress signals or SOS sounds emitted user. By integrating advanced sensing technologies intelligent algorithms, aims provide prompt assistance ensure safety passengers transportation settings.

Language: Английский

Citations

0
Bioinspired Garra Rufa Optimization-Assisted Deep Learning Model for Object Classification on Pedestrian Walkways DOI Creative Commons
Eunmok Yang, K. Shankar, Sachin Kumar

et al.

Biomimetics, Journal Year: 2023, Volume and Issue: 8(7), P. 541 - 541

Published: Nov. 11, 2023

Object detection in pedestrian walkways is a crucial area of research that widely used to improve the safety pedestrians. It not only challenging but also tedious process manually examine labeling abnormal actions, owing its broad applications video surveillance systems and larger number videos captured. Thus, an automatic system identifies anomalies has become indispensable for computer vision (CV) researcher workers. The recent advancements deep learning (DL) algorithms have attracted wide attention CV processes such as object classification based on supervised requires labels. current study designs bioinspired Garra rufa optimization-assisted model (BGRODL-OC) technique walkways. objective BGRODL-OC recognize presence pedestrians objects video. To achieve this goal, primarily applies GhostNet feature extractors produce set vectors. In addition this, makes use GRO algorithm hyperparameter tuning process. Finally, performed via attention-based long short-term memory (ALSTM) network. A range experimental analysis was conducted validate superior performance technique. values established over other existing approaches.

Language: Английский

Citations

1