Road infrastructure maintenance in developing countries faces unique challenges due to resource constraints and diverse environmental factors. This study addresses the critical need for efficient, accurate, and locally-relevant pavement distress detection methods in these regions. We present a novel deep learning approach combining YOLO (You Only Look Once) object detection models with a Convolutional Block Attention Module (CBAM) to simultaneously detect and classify multiple pavement distress types. The model demonstrates robust performance in detecting and classifying potholes, longitudinal cracks, alligator cracks, and raveling, with confidence scores ranging from 0.46 to 0.93. While some misclassifications occur in complex scenarios, these provide insights into unique challenges of pavement assessment in developing countries. Additionally, we developed a web-based application for real-time distress detection from images and videos. This research advances automated pavement distress detection and provides a tailored solution for developing countries, potentially improving road safety, optimizing maintenance strategies, and contributing to sustainable transportation infrastructure development.

Object detection in images or videos faces several challenges because the detection must be accurate, efficient and fast. The you only look once (YOLO) algorithm was invented to meet these criteria. But with the creation of several versions of this algorithm (from V1 to V11), it becomes difficult for researchers to choose the best one. The main objective of this review is to present and compare the eleven versions of the yolo algorithm in order to know when using the appropriate one for the study. The methodology used for this work is aligned with preferred reporting items for systematic reviews and meta-analyses (PRISMA) principles and the results demonstrate that the choice of the best version mainly depends on the priorities of the study. If the study prioritizes accuracy and detection of small objects, it should use YOLO V4, YOLO V5, YOLO V6, YOLO V7, YOLO V8, YOLO V9, YOLO V10 or YOLO V11. While studies that prioritize detection speed should use YOLO V5, YOLO V6, YOLO V7, YOLO V8, YOLO V10 or YOLO V11. In complex environment, researchers should avoid using YOLO V1, YOLO V2, YOLO V3, YOLO V5, YOLO V7 and YOLO V9. And researchers who are looking for a good accuracy and speed and a reduced number of parameters should use YOLO V10 or YOLO V11.

Boron Neutron Capture Therapy (BNCT) is an advanced form of radiotherapy that uses the neutron capture reaction of boron-10 to selectively destroy cancer cells. Accurate and reproducible patient positioning is critical to treatment efficacy, yet conventional workflows rely on manual adjustments and laser alignment, introducing operator dependence and potential geometric uncertainty. This study presents an integrated Surface-Guided BNCT patient positioning system (SG-BNCT) that combines a binocular stereo-vision (BSV) module, the BNCT-specific treatment planning system NeuMANTA, and a six-axis industrial robot. The BSV system uses fiducial markers with You Only Look Once (YOLO)-based detection and stereo triangulation to reconstruct 3D geometry, while TPS-derived transformation matrices drive six-degree-of-freedom robotic adjustments. Validation with anthropomorphic phantoms demonstrated millimeter accuracy. By eliminating laser dependence and providing closed-loop corrections, SG-BNCT enhances positioning precision, reduces operator variability, and streamlines setup, supporting more reliable and efficient BNCT treatment delivery.

To maintain reliability, safety, and sustainability in power transmission, insulator defect detection has become a critical task in power line inspection. Due to the complex backgrounds and small defect sizes encountered in insulator defect images, issues such as false detections and missed detections often occur. The existing You Only Look Once (YOLO) object detection algorithm is currently the mainstream method for image-based insulator defect detection in power lines. However, existing models suffer from low detection accuracy. To address this issue, this paper presents an improved YOLOv5-based MC-YOLO insulator detection algorithm. To effectively extract multi-scale information and enhance the model’s ability to represent feature information, a multi-scale attention convolutional fusion (MACF) module incorporating an attention mechanism is proposed. This module utilises parallel convolutions with different kernel sizes to effectively extract features at various scales and highlights the feature representation of key targets through the attention mechanism, thereby improving the detection accuracy. Additionally, a cross-context feature fusion module (CCFM) is designed, where shallow features gain partial deep semantic supplementation and deep features absorb shallow spatial information, achieving bidirectional information flow. Furthermore, the Spatial-Channel Dual Attention Module (SCDAM) is introduced into CCFM. By incorporating a dynamic attention-guided bidirectional cross-fusion mechanism, it effectively resolves the feature deviation between shallow details and deep semantics during multi-scale feature fusion. The experimental results show that the MC-YOLO algorithm achieves an mAP@0.5 of 67.4% on the dataset used in this study, which is a 4.1% improvement over the original YOLOv5. Although the FPS is slightly reduced compared to the original model, it remains practical and capable of rapidly and accurately detecting insulator defects.

The rapid growth of intelligent environmental security systems has intensified the need for accurate and real-time suspicious human activity recognition. Computer vision techniques, particularly deep learning–based object detection models, have emerged as key enablers in addressing these challenges. Among them, You Only Look Once (YOLO) has gained significant attention due to its high detection speed, end-to-end architecture, and suitability for real-time surveillance applications. This review paper presents a comprehensive analysis of the application of YOLO-based models in suspicious human activity recognition for intelligent environmental security systems. It examines the evolution of YOLO architectures, their adaptations for activity and behavior analysis, and their integration with surveillance frameworks. The review further discusses commonly used datasets, performance evaluation metrics, and comparative results reported in existing studies. In addition, key challenges such as occlusion, varying illumination, complex backgrounds, privacy concerns, and computational constraints are highlighted. Finally, the paper outlines future research directions, including hybrid models, multi-modal data fusion, edge-based deployment, and explainable AI, to enhance the robustness and reliability of YOLO-driven security systems. This review aims to provide researchers and practitioners with a structured understanding of current advancements and open issues in YOLO-based suspicious human activity recognition.

Thermal comfort and energy efficiency are two main goals of heating, ventilation, and air conditioning (HVAC) systems, which use about 40% of the total energy in buildings. This paper aims to predict optimal room temperature, enhance comfort, and reduce energy consumption while avoiding extra energy use from overheating or overcooling. Six Machine Learning (ML) models were tested to predict the optimal temperature in the classroom based on the occupancy characteristic detected by a Deep Learning (DL) model, You Only Look Once (YOLO). The decision tree achieved the highest accuracy at 97.36%, demonstrating its effectiveness in predicting the preferred temperature. To measure energy savings, the study used RETScreen software version 9.4 to compare intelligent temperature control with traditional operation of HVAC. Genetic algorithm (GA) was further employed to optimize HVAC energy consumption while keeping the thermal comfort level by adjusting set-points based on real-time occupancy. The GA showed how to balance comfort and efficiency, leading to better system performance. The results show that adjusting from default HVAC settings to preferred thermal comfort levels as well controlling the HVAC to work only if the room is occupied can reduce energy consumption and costs by approximately 76%, highlighting the substantial impact of even simple operational adjustments. Further improvements achieved through GA-optimized temperature settings provide additional savings of around 7% relative to preferred comfort levels, demonstrating the value of computational optimization techniques in fine-tuning building performance. These results show that intelligent, data-driven HVAC control can improve comfort, save energy, lower costs, and support sustainability in buildings.

Timely and accurate detection of forest fires through unmanned aerial vehicle (UAV) remote sensing target detection technology is of paramount importance. However, multiscale targets and complex environmental interference in UAV remote sensing images pose significant challenges during detection tasks. To address these obstacles, this paper presents FF-Mamba-YOLO, a novel framework based on the principles of Mamba and YOLO (You Only Look Once) that leverages innovative modules and architectures to overcome these limitations. Specifically, we introduce MFEBlock and MFFBlock based on state space models (SSMs) in the backbone and neck parts of the network, respectively, enabling the model to effectively capture global dependencies. Second, we construct CFEBlock, a module that performs feature enhancement before SSM processing, improving local feature processing capabilities. Furthermore, we propose MGBlock, which adopts a dynamic gating mechanism, enhancing the model’s adaptive processing capabilities and robustness. Finally, we enhance the structure of Path Aggregation Feature Pyramid Network (PAFPN) to improve feature fusion quality and introduce DySample to enhance image resolution without significantly increasing computational costs. Experimental results on our self-constructed forest fire image dataset demonstrate that the model achieves 67.4% mAP@50, 36.3% mAP@50:95, and 64.8% precision, outperforming previous state-of-the-art methods. These results highlight the potential of FF-Mamba-YOLO in forest fire monitoring.

The fast growth of online banking and mobile financial services has heightened the demand for secure, easy-to-use authentication mechanisms. Traditional methods like passwords and one-time passwords are increasingly susceptible to cyber-attacks, which constitutes the main motivation towards the adoption of biometric-based authentications. From biometric modalities, face recognition has gained widespread acceptance owing to its non-intrusive nature and suitability for mobile devices. Recent deep learning advancements have made real-time face detection and recognition possible via object detection frameworks like YOLO (You Only Look Once). This work presents a comprehensive analysis of a YOLO-based biometric authentication system devised for online banking and mobile applications. This paper proposes a complete biometric pipeline that uses YOLO for face detection and deep embedding-based models for recognition. Face detection performance is evaluated on the WIDER FACE dataset, while recognition accuracy is assessed on the LFW dataset. The paper presents a reproducible implementation in detail through a Google Colab environment. System performance is analyzed in terms of detection accuracy, recognition accuracy, inference speed, and end-to-end latency. An extensive ablation study investigates the impact of key components, including detection architectures, face alignment strategies, embedding model selection, and similarity threshold tuning. Furthermore, the proposed research framework is compared against Zoloz, a commercial enterprise-grade biometric authentication platform widely adopted in the banking sector. The results show that YOLO-based biometric systems are very effective for research and prototyping, while real-world banking deployment requires additional security, compliance, and robustness considerations.

Considering the rapid iteration of you-only-look-once (YOLO)-series algorithms, this paper aims to provide a data-driven performance spectrum and selection guide for the latest YOLO series algorithm (YOLOv8 to YOLOv13) in printed circuit board (PCB) automatic optical inspection (AOI) through systematic benchmarking. A comprehensive evaluation of the six state-of-the-art YOLO series algorithms is conducted on a standardized dataset containing six typical PCB defects: missing hole, mouse bite, open circuit, short circuit, spur, and spurious copper. An innovative dual-cycle comparative experiment (100 rounds and 500 rounds) is designed, and a systematic assessment is performed across multiple dimensions, including accuracy, efficiency, and inference speed. The experimental results have revealed significant variations in algorithm performance with training cycles: under short-term training (100 rounds), YOLOv13 achieves leading detection performance (mAP50 = 0.924, mAP50-95 = 0.484) with the fewest parameters (2.45 million); after full training (500 rounds), YOLOv10 achieves the highest overall accuracy (mAP50 = 0.946, mAP50-95 = 0.526); additionally, YOLOv11 shows the optimal speed-accuracy balance after long-term training, while YOLOv12 excels in short-term training; moreover, “open circuit” and “spur” are evaluated as the most challenging defect categories to detect. The findings given in this paper indicate the absence of a universally applicable “all-in-one” algorithm and propose a clear algorithm selection roadmap: YOLOv10 is recommended for offline analysis scenarios prioritizing extreme accuracy; YOLOv13 is the top choice for applications requiring rapid iteration with tight training time constraints; and YOLOv11 is the best option for high-throughput online inspection PCB production lines.

The aim of this study was to develop an artificial intelligence (AI) model capable of quantifying crowding and spacing in the upper arch and to validate its accuracy by comparing the model's results with those of manual interactive digital space analysis. This study included intraoral photographs and occlusal scans of the upper dental arch of orthodontic patients treated at the University of Sharjah (2022-2024). The YOLO (You Only Look Once) 8Pose Model was generated using atraining and validation dataset (832 images). The AI model performed tooth segmentation and tooth point detection on the occlusal images, followed by automated quantification of tooth size-arch length discrepancy (TSALD). Manual space analysis was conducted using OrthoCAD (Cadent, Fairview, NJ, USA) software and the data were compared with the results of the AI model using atesting dataset (300 images). TSALD was categorized based on the index of treatment complexity, outcome, and need (ICON). Qualitative data were presented as frequency and distribution, and comparisons were performed by using Fisher's exact test. Correlation between manual and AI-measured TSALD was evaluated using Pearson's correlation coefficient. The model achieved an overall accuracy of 90%. The largest discrepancies were found in cases presenting mild crowding (< 2 mm, 7%), severe spacing (5.1-9 mm, 5%), and moderate spacing (2.1-5 mm, 3.3%). Astrong correlation (> 0.92) between manual and AI TSALD measurements indicated high reliability and potential interchangeability. The AI model was successfully developed and validated, achieving 90% accuracy, demonstrating its potential as areliable tool for quantifying TSALD in orthodontic diagnostics.

AI-powered detection of dental anatomy: a YOLO-based approach.

YOLO-spectra: A generalized framework for rapid simultaneous detection and classification of Raman spectra in images with mobile devices for enhancing on-site applications.

AMSA-YOLO: Real-time object detection with adaptive multi-scale attention mechanism.

Lightweight frameworks for real-time crack monitoring in civil infrastructure.

Rapid and on-site detection of carbaryl in complex matrices using cellulose-based organogel sensor with assist of machine learning.

Existing infrared imaging techniques have garnered considerable attention and have achieved notable progress in all weather ship target detection tasks, owing to their robustness against varying ambient lighting conditions. However, due to the inherent limitations of infrared images,such as low spatial resolution and insufficient texture information the performance of multi-scale ship target detection remains suboptimal. These challenges significantly hinder the overall improvement of detection accuracy. To address this issue and enhance the detection performance of multi-scale ship targets, particularly small ones, in infrared imagery,this paper proposes an improved You Only Look Once (YOLO) based detection model named ADN-YOLO. The model first introduces a Dynamic Upsampler (Dysample) module, which more effectively integrates semantic information across different layers. This integration balances low level detailed features with high level semantic representations, thereby enhancing the model?s ability to perceive target edges and structural characteristics. Second, a lightweight downsampling module (ADown) is incorporated to reduce the parameter count while improving both the efficiency and representational capacity of feature extraction. Additionally, to address the issue of small targets being highly sensitive to localization errors, a new loss function is designed based on the Wasserstein distance. This function combines the Normalized Wasserstein Distance (NWD) with the Complete Intersection over Union (CIoU), thereby enhancing the model?s ability to accurately localize small targets. Comprehensive experimental validation is conducted on a marine infrared target detection dataset. Compared to the standard YOLOv11 model, the proposed ADN-YOLO reduces the number of parameters by 20.3%, achieves a 1.9%increase in mAP, a 1.9% boost in Recall, and lowers FLOPs by 1.1G, demonstrating its effectiveness and practicality for infrared image target detection tasks.

ABSTRACT Fused Deposition Modelling (FDM) is the predominant material extrusion technique in polymer additive manufacturing (AM). While it offers compatibility with engineering-grade composites and enables the fabrication of polymer-composite components with intricate architectures unattainable through traditional techniques, the persistent dependence on empirical process tuning often leads to structural defects – critical limitations that hinder FDM's transition to advanced industrial applications. This paper proposes a data-driven approach that integrates advanced Artificial Intelligence (AI) with real-time computer vision to optimise FDM process parameters and enable in-process anomaly detection. In the developed approach, a stepwise machine learning strategy systematically models the printed line quality, ensuring pre-print process optimisation. A You Only Look Once (YOLO) object detection model is then deployed for in-situ monitoring, analysing the printed line morphology to assess melt flow stability and detect geometric deviations during printing. Validation experiments are conducted to assess the effectiveness of the developed YOLO model. Overall, the integrated framework demonstrates its superiority over empirical methods and analytical models in both pre-process optimisation and real-time quality assurance. Furthermore, the integrated machine vision and pattern recognition system exhibits adaptability to diverse material deposition systems, providing a unified approach to intelligent process optimisation across AM domains.

Abstract The detection of surface defects in industry is of great importance in ensuring the quality of industrial products. There are low-quality examples in defect detection that are commonly found in small and extreme aspect ratio targets. How to improve the performance of algorithms for industrial surface defect detection, especially in enhancing the detection ability of small-sized and extreme aspect ratio targets while ensuring inference speed, has not been deeply studied at present. Firstly, in response to the challenge of difficulty in detecting multiscale objects in industrial surface defects, this paper proposes the cross-level multiscale feature fusion (CMFF) module that integrates cross-level features from both the backbone and neck. Secondly, to alleviate the issue of low-quality examples leading to difficult localization, this paper proposes a novel regression loss function, by reweighting low-quality examples to guide the model towards better localization. Finally, the cross-level multiscale feature fusion enhanced You Only Look Once (YOLO) with focused-Complete Intersection over Union (CIoU) loss, named CMFF-YOLO network is proposed. By integrating the CMFF module and the focused-CIOU loss function, the proposed network is used for defect detection. The CMFF-YOLO model achieved mean average precision (mAP) values of 84.2%, 76.6%, and 94.8% on the NEU-DET, GC10-DET, and PCB-DET datasets, respectively, outperforming the baseline YOLOv8s by 3.4%, 6.6%, and 3.0%. Among the 12 models evaluated, this method ranked among the top three for inference speed.

Horticultural agriculture, especially guava (Psidium guajava), has great economic potential in Indonesia. However, productivity often declines due to fruit disease attacks, which are still manually diagnosed by farmers. This study aims to develop an artificial intelligence-based guava disease classification system using the You Only Look Once (YOLO) version 5 algorithm. The dataset consists of 600 images divided into three disease classes: Phytophthora, Styler and Root, and Scab. Data were collected through field documentation, then preprocessed and augmented using Roboflow. The dataset was divided into 70% training data, 20% validation, and 10% testing. The YOLOv5 model was trained using Google Collaboratory and consistently evaluated using the Confusion Matrix and accuracy, precision, recall, and F1-score metrics. The test results showed that the model achieved an accuracy of more than 95% with high precision, recall, and F1-score values ​​for each disease class. This proves that YOLOv5 is effective for real-time guava disease detection. This research contributes to the application of artificial intelligence technology to help farmers make early diagnoses quickly and accurately, thereby reducing the risk of reduced crop yields.

YOLO (You Only Look Once) is a one-stage detector that predicts object classes and bounding boxes in a single pass without an explicit region proposal step. In contrast, two-stage detectors first generate candidate regions. The YOLO NAS-L model is specifically designed to improve the detection of small objects. The purpose of this study is to systematically investigate the influence of dataset characteristics, training strategies and hyperparameter selection on the performance of YOLO NAS-L in a challenging object detection scenario: detecting swimmers in aquatic environments. Using both the mean Average Precision value (mAP)—which reflects the model’s global precision–recall performance and the F1-score, indicating the model’s effectiveness under realistic operating conditions—as evaluation metrics, this study investigates the effects of batch size, batch accumulation, number of training epochs, image resolution, pre-trained weights, and data augmentation. Our findings indicate that while batch size and image resolution had limited impact on performance parameters, the use of batch accumulation, pre-trained weights and careful tuning of training epochs were critical for optimizing model performance. The results highlight the practical significance of combining optimized hyperparameters, training strategies, and pre-trained weights to efficiently develop high-performing YOLO NAS-L models.