Machine Learning Algorithms for Manufacturing Quality Assurance: A Systematic Review of Performance Metrics and Applications

Array, Journal Year: 2025, Volume and Issue: unknown, P. 100393 - 100393

Published: March 1, 2025

Language: Английский

---

Extraction of Solar Panel Image Texture Feature Using GLCM Method for Damage Analysis on Solar Panel Surface Images

E3S Web of Conferences, Journal Year: 2025, Volume and Issue: 622, P. 01001 - 01001

Published: Jan. 1, 2025

Existing techniques for assessing solar panel surface damage frequently lack precision in differentiating defect kinds, necessitating a dependable automated solution. Defects like cracks and scratches substantially diminish efficiency, underscoring the necessity of robust analytical procedures. This study seeks to validate Gray Level Cooccurrence Matrix (GLCM) technique extracting texture information identify analyze on surfaces. method utilizes Python software dataset photos accurately distinguish between damaged undamaged The spot category demonstrates lowest homogeneity (5636.922) contrast (5632.922), signifying smoother yet less uniform texture. Energy values are predominantly low across all categories, with marginally higher consistency fractures (0.005) relative others (0.002). results indicate that faults enhance unpredictability randomization intact These insights facilitate precise identification enhanced maintenance plans. research provides advancements renewable energy, materials science, computer vision, applicable maintenance, quality assurance, flaw within photovoltaic sector.

Language: Английский

---

Real-time structural health monitoring of bridges using convolutional neural network-based loss factor analysis for enhanced energy dissipation detection

Structures, Journal Year: 2024, Volume and Issue: 70, P. 107733 - 107733

Published: Nov. 12, 2024

Language: Английский

---

Machine Learning in 3D and 4D Printing of Polymer Composites: A Review

Polymers, Journal Year: 2024, Volume and Issue: 16(22), P. 3125 - 3125

Published: Nov. 8, 2024

The emergence of 3D and 4D printing has transformed the field polymer composites, facilitating fabrication complex structures. As these manufacturing techniques continue to progress, integration machine learning (ML) is widely utilized enhance aspects processes. This includes optimizing material properties, refining process parameters, predicting performance outcomes, enabling real-time monitoring. paper aims provide an overview recent applications ML in composites. By highlighting intersection technologies, this seeks identify existing trends challenges, outline future directions.

Language: Английский

---

Deep learning-based anomaly detection using one-dimensional convolutional neural networks (1D CNN) in machine centers (MCT) and computer numerical control (CNC) machines

PeerJ Computer Science, Journal Year: 2024, Volume and Issue: 10, P. e2389 - e2389

Published: Oct. 17, 2024

Computer numerical control (CNC) and machine center (MCT) machines are mechanical devices that manipulate different tools using computer programming as inputs. Predicting failures in CNC MCT before their actual failure time is crucial to reduce maintenance costs increase productivity. This study centered around a novel deep learning-based model 1D convolutional neural network (CNN) for early fault detection machines. We collected sensor-based data from CNC/MCT applied various preprocessing techniques prepare the dataset. Our experimental results demonstrate 1D-CNN achieves higher accuracy of 91.57% compared traditional learning classifiers other models, including Random Forest (RF) at 89.71%, multi-layer perceptron (MLP) 87.45%, XGBoost 89.67%, logistic regression (LR) 75.93%, support vector (SVM) 75.96%, K-nearest neighbors (KNN) 82.93%, decision tree 88.36%, naïve Bayes 68.31%, long short-term memory (LSTM) 90.80%, hybrid CNN + LSTM 88.51%. Moreover, our proposed outperformed all mentioned models precision, recall, F-1 scores, with 91.87%, 91.57%, 91.63%, respectively. These findings highlight efficacy providing optimal performance an machine’s dataset, making it particularly suitable small manufacturing companies seeking automate classification approach enhances productivity aids proactive safety measures, demonstrating its potential revolutionize industry.

Language: Английский

---

Hierarchical Image Quality Improvement Based on Illumination, Resolution, and Noise Factors for Improving Object Detection

Electronics, Journal Year: 2024, Volume and Issue: 13(22), P. 4438 - 4438

Published: Nov. 12, 2024

Object detection performance is significantly impacted by image quality factors such as illumination, resolution, and noise. This paper proposes a hierarchical improvement process that dynamically prioritizes these based on severity, enhancing accuracy in diverse conditions. The evaluates each factor—illumination, noise—using discriminators analyze brightness, edge strength, noise levels. Improvements are applied iteratively with an adaptive weight update mechanism adjusts factor importance effectiveness. Following improvement, assessment conducted, updating weights to fine-tune subsequent adjustments. allows the learn optimal parameters for varying conditions, adaptability. improved through proposed shows index (PSNR, SSIM) evaluation, object when measured using deep learning models called YOLOv8 RT-DETR. rate 7% ‘Bottle’ high-light environment, 4% 2.5% ‘Bicycle’ ‘Car’ objects low-light respectively. Additionally, segmentation saw 9.45% gain, supporting effectiveness of this method real-world applications.

Language: Английский

---

Biologically inspired oscillating activation functions can bridge the performance gap between biological and artificial neurons

Expert Systems with Applications, Journal Year: 2024, Volume and Issue: unknown, P. 126036 - 126036

Published: Dec. 1, 2024

Language: Английский

---

Artificial Intelligence-Driven Advances in Coronary Calcium Scoring: Expanding Preventive Cardiology

Cureus, Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 28, 2024

Coronary artery disease (CAD) remains a leading global cause of morbidity and mortality, underscoring the need for effective cardiovascular risk stratification preventive strategies. calcium (CAC) scoring, traditionally performed using electrocardiogram (ECG)-gated cardiac computed tomography (CT) scans, has been widely validated as robust tool assessing risk. However, its application largely limited to high-risk populations due costs, technical requirements, accessibility CT scans. Recent advancements in artificial intelligence (AI) have introduced transformative opportunities extend CAC detection noncardiac such those lung cancer screening, enabling broader more accessible screening. This review provides comprehensive analysis AI-driven detection, examining various types AI models like convolutional neural networks (CNNs) U-Net architectures, exploring clinical, operational, ethical implications incorporating these technologies into routine practice. Technical challenges, including imaging variability, data privacy, model bias, are discussed alongside essential areas further research, standardization validation across diverse populations. By leveraging available data, AI-enabled potential advance cardiology, supporting earlier identification, optimizing healthcare resources, improving patient outcomes.

Language: Английский

---

Design and Development of a Precision Defect Detection System Based on a Line Scan Camera Using Deep Learning

Applied Sciences, Journal Year: 2024, Volume and Issue: 14(24), P. 12054 - 12054

Published: Dec. 23, 2024

The manufacturing industry environment is rapidly evolving into smart manufacturing. It prioritizes digital innovations such as AI and transformation (DX) to increase productivity create value through automation intelligence. Vision systems for defect detection quality control are being implemented across industries, including electronics, semiconductors, printing, metal, food, packaging. Small medium-sized companies increasingly demanding factory solutions added enhance competitiveness. In this paper, we design develop a high-speed system based on line-scan camera using deep learning. positioned side-view imaging, allowing detailed inspection of the component mounting soldering PCBs. To detect defects PCBs, gathers extensive images both flawless defective products train learning model. An engine generated process then applied conduct inspections. developed was evaluated have an accuracy 99.5% in experiment. This will be highly beneficial precision management small- enterprises

Language: Английский

---

Enhanced convolutional neural network methodology for solid waste classification utilizing data augmentation techniques

Waste Management Bulletin, Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 1, 2024

Language: Английский

---