Traffic Management: Multi-Scale Vehicle Detection in Varying Weather Conditions Using YOLOv4 and Spatial Pyramid Pooling Network

Electronics, Journal Year: 2022, Volume and Issue: 11(17), P. 2748 - 2748

Published: Sept. 1, 2022

Detecting and counting on road vehicles is a key task in intelligent transport management surveillance systems. The applicability lies both urban highway traffic monitoring control, particularly difficult weather conditions. In the past, has been performed through data acquired from sensors conventional image processing toolbox. However, with advent of emerging deep learning based smart computer vision systems become computationally efficient reliable. mounted cameras can be used to train models which detect track for analysis handling problems such as congestion harsh conditions where there are poor visibility issues because low illumination blurring. Different vehicle detection algorithms focusing same issue deal only or two specific this research, we address detecting scene multiple scenarios including haze, dust sandstorms, snowy rainy day nighttime. proposed architecture uses CSPDarknet53 baseline modified spatial pyramid pooling (SPP-NET) layer reduced Batch Normalization layers. We also augment DAWN Dataset different techniques Hue, Saturation, Exposure, Brightness, Darkness, Blur Noise. This not increases size dataset but make more challenging. model obtained mean average precision 81% during training detected smallest present

Language: Английский

---

YOLO-Based Deep Learning Model for Pressure Ulcer Detection and Classification

Healthcare, Journal Year: 2023, Volume and Issue: 11(9), P. 1222 - 1222

Published: April 25, 2023

Pressure ulcers are significant healthcare concerns affecting millions of people worldwide, particularly those with limited mobility. Early detection and classification pressure crucial in preventing their progression reducing associated morbidity mortality. In this work, we present a novel approach that uses YOLOv5, an advanced robust object model, to detect classify into four stages non-pressure ulcers. We also utilize data augmentation techniques expand our dataset strengthen the resilience model. Our shows promising results, achieving overall mean average precision 76.9% class-specific mAP50 values ranging from 66% 99.5%. Compared previous studies primarily CNN-based algorithms, provides more efficient accurate solution for The successful implementation has potential improve early treatment ulcers, resulting better patient outcomes reduced costs.

Language: Английский

---

Explainable AI in Healthcare Application

Advances in computational intelligence and robotics book series, Journal Year: 2024, Volume and Issue: unknown, P. 123 - 176

Published: Jan. 18, 2024

Given the inherent risks in medical decision-making, professionals carefully evaluate a patient's symptoms before arriving at plausible diagnosis. For AI to be widely accepted and useful technology, it must replicate human judgment interpretation abilities. XAI attempts describe data underlying black-box approach of deep learning (DL), machine (ML), natural language processing (NLP) that explain how judgments are made. This chapter provides survey most recent methods employed imaging related fields, categorizes lists types XAI, highlights used make topics more interpretable. Additionally, focuses on challenging issues applications guides development better deep-learning system explanations by applying principles analysis pictures text.

Language: Английский

---

Use of Deep Learning Applications for Drone Technology

Advances in information security, privacy, and ethics book series, Journal Year: 2024, Volume and Issue: unknown, P. 128 - 147

Published: Jan. 26, 2024

Imagine a society where conventional techniques no longer constrain crime investigation and instead use cutting-edge technology to crack cases more quickly effectively. With the development of deep learning drone technology, this is world we are heading towards. Investigators may now collect critical evidence from previously inaccessible sites analyse it with extraordinary accuracy because combination these two fields. There tremendous promise for solving crimes believed be unsolvable, ramifications justice significant. Drones, often referred as unmanned aerial vehicles (UAVs), becoming increasingly widespread in various settings, including businesses, factories, leisure. However, due their growing popularity, there worries about drone-related crime.

Language: Английский

---

Zero-Day Exploits in Cybersecurity: Case Studies and Countermeasure

Published: July 29, 2024

Zero-day threats are a more severe and constantly developing menace to various participants including large companies, government offices, educational establishments. These entities may contain valuable information essential operations that attract cyber attackers. exploits especially devastating as they target weaknesses an organization’s vendors not even aware of, making them have no protection against them. This paper focuses on the background use of zero-day exploitation structure technologies these complex malware attacks. We examine two notable real-life cases: case ‘HAFNIUM targeting Exchange Servers with exploits’ was investigated by Microsoft 365 Security Threat Intelligence, ‘Log4j vulnerability’ reported National Cyber Centre. cases show critical effects vulnerabilities measures taken combat Additionally, this outlines different strategies can be used prevent attacks help modern technologies. fast patch release, effective IDS/IPS, security model involves constant vigilance behavioral analytics. Thus, studying lifecycle exploits, one enhance organization invisible traditional systems. extensive survey is designed useful in understanding characteristics vulnerabilities, for their mitigation, threat development field cybersecurity. it possible strengthen develop time analyzing previous events predicting potential problems.

Language: Английский

---

COVID-19 Hierarchical Classification Using a Deep Learning Multi-Modal

Sensors, Journal Year: 2024, Volume and Issue: 24(8), P. 2641 - 2641

Published: April 20, 2024

Coronavirus disease 2019 (COVID-19), originating in China, has rapidly spread worldwide. Physicians must examine infected patients and make timely decisions to isolate them. However, completing these processes is difficult due limited time availability of expert radiologists, as well limitations the reverse-transcription polymerase chain reaction (RT-PCR) method. Deep learning, a sophisticated machine learning technique, leverages radiological imaging modalities for diagnosis image classification tasks. Previous research on COVID-19 encountered several limitations, including binary methods, single-feature modalities, small public datasets, reliance CT diagnostic processes. Additionally, studies have often utilized flat structure, disregarding hierarchical structure pneumonia classification. This study aims overcome by identifying caused COVID-19, distinguishing it from other types healthy lungs using chest X-ray (CXR) images related tabular medical data, demonstrate value incorporating data achieving more accurate diagnoses. Resnet-based VGG-based pre-trained convolutional neural network (CNN) models were employed extract features, which then combined early fusion eight distinct classes. We leveraged hierarchal within our approach achieve improved outcomes. Since an imbalanced dataset common this field, variety versions generative adversarial networks (GANs) used generate synthetic data. The proposed tested private datasets 4523 achieved macro-avg F1-score 95.9% 87.5% identification structure. In conclusion, study, we able create deep multi-modal diagnose differentiate kinds normal lungs, will enhance process.

Language: Английский

---

Framework for Improved Sentiment Analysis via Random Minority Oversampling for User Tweet Review Classification

Electronics, Journal Year: 2022, Volume and Issue: 11(19), P. 3058 - 3058

Published: Sept. 25, 2022

Social networks such as twitter have emerged social platforms that can impart a massive knowledge base for people to share their unique ideas and perspectives on various topics issues with friends families. Sentiment analysis based machine learning has been successful in discovering the opinion of using redundantly available data. However, recent studies pointed out imbalanced data negative impact results. In this paper, we propose framework improved sentiment through ordered preprocessing steps combination resampling minority classes produce greater performance. The performance technique vary depending dataset its initial focus is feature selection combination. Multiple algorithms are utilized classification tweets into positive, negative, or neutral. Results revealed random oversampling provide it tackle issue class imbalance.

Language: Английский

---

Religious Sentiment Analysis and Detection on Social Media

Lecture notes in networks and systems, Journal Year: 2025, Volume and Issue: unknown, P. 245 - 262

Published: Jan. 1, 2025

Language: Английский

---

An Implementation of RFID-Enabled ID Cards in the Education Sector to Optimize and Facilitate the Use of Student Cards

Lecture notes in networks and systems, Journal Year: 2025, Volume and Issue: unknown, P. 93 - 107

Published: Jan. 1, 2025

Language: Английский

---

Challenges and Limitations in Data Mining

Lecture notes in networks and systems, Journal Year: 2025, Volume and Issue: unknown, P. 207 - 220

Published: Jan. 1, 2025

Language: Английский

---