Evaluating the Effectiveness of YOLO Models in Different Sized Object Detection and Feature-Based Classification of Small Objects

Quach, Luyl-Da; Quoc, Khang Nguyen; Quynh, Anh Nguyen; Ngoc, Hoang Tran

doi:10.12720/jait.14.5.907-917

Cited by 26 publications

(2 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The network structure of Yolov5 followed the design principles of the Yolo [8][9] series, which consists of four key components: input, Backbone, Neck, and prediction result output. To expedite model convergence, Mosaic data augmentation is applied at the input stage.…”

Section: ░ 2 Deep Learning Yolov5 Algorithmmentioning

confidence: 99%

Advancements in Steel Surface Defect Detection: An Enhanced YOLOv5 Algorithm with EfficientNet Integration

Ren,

Zhang,

Fei

et al. 2024

IJEER

View full text Add to dashboard Cite

Steel surface defect detection is of utmost importance for ensuring product quality, cost reduction, enhanced safety, and heightened customer satisfaction. To address the limitations of traditional steel surface defect detection algorithms, which often yielded singular detection results and suffered from high miss detection rates, we proposed an enhanced Yolov5 steel surface defect detection algorithm. In this approach, this paper employed the EfficientNet network as a replacement for the Yolov5 backbone network. Subsequently, we trained and tested this modified network on a steel surface defect dataset to mitigate the challenges associated with high miss detection rates and underperforming evaluation metrics. Our experimental findings underscored the superiority of the improved algorithm, particularly when compared to Yolov5. This enhanced algorithm exhibited substantial improvements across several key performance metrics, including Precision, Recall, mAP@0.5, parameter count, and pt file size. Noteworthy achievements included a 6.39% increase in Precision for Yolov5-EfficientNetB4, a remarkable 7.75% improvement in Recall for Yolov5-EfficientNetB0, and a 5.57% boost in mAP@0.5 for Yolov5-EfficientNetB6. Additionally, the pt file size for Yolov5-EfficientNetB0 saw a substantial 39.65% reduction, although it was important to note that the inference time for the improved algorithm increased. Among the models, Yolov5-EfficientNetB6 struck the best balance in terms of performance.

show abstract

Section: ░ 2 Deep Learning Yolov5 Algorithmmentioning

confidence: 99%

Advancements in Steel Surface Defect Detection: An Enhanced YOLOv5 Algorithm with EfficientNet Integration

Ren,

Zhang,

Fei

et al. 2024

IJEER

View full text Add to dashboard Cite

show abstract

“…In Equation (11), n is the threshold level belonging to real numbers, and the values are in the range of 0 to 1, whereas N denotes the total number of classes. The average precision (AP) and mAP values are fundamental for object detection, offering a comprehensive view of the ability of an algorithm to identify and localize objects accurately within images [88], accounting for precision and recall trade-offs [89].…”

Section: Evaluation Metricsmentioning

confidence: 99%

Vision-Based Construction Safety Monitoring Utilizing Temporal Analysis to Reduce False Alarms

Zaidi,

Yang,

Abbas

et al. 2024

Buildings

View full text Add to dashboard Cite

Construction safety requires real-time monitoring due to its hazardous nature. Existing vision-based monitoring systems classify each frame to identify safe or unsafe scenes, often triggering false alarms due to object misdetection or false detection, which reduces the overall monitoring system’s performance. To overcome this problem, this research introduces a safety monitoring system that leverages a novel temporal-analysis-based algorithm to reduce false alarms. The proposed system comprises three main modules: object detection, rule compliance, and temporal analysis. The system employs a coordination correlation technique to verify personal protective equipment (PPE), even with partially visible workers, overcoming a common monitoring challenge on job sites. The temporal-analysis module is the key component that evaluates multiple frames within a time window, triggering alarms when the hazard threshold is exceeded, thus reducing false alarms. The experimental results demonstrate 95% accuracy and an F1-score in scene classification, with a notable 2.03% average decrease in false alarms during real-time monitoring across five test videos. This study advances knowledge in safety monitoring by introducing and validating a temporal-analysis-based algorithm. This approach not only improves the reliability of safety-rule-compliance checks but also addresses challenges of misdetection and false alarms, thereby enhancing safety management protocols in hazardous environments.

show abstract

Grain rot dataset caused by Burkholderia Glumae Bacteria

Quoc,

Quach

2024

Data in Brief

View full text Add to dashboard Cite

Evaluating the Effectiveness of YOLO Models in Different Sized Object Detection and Feature-Based Classification of Small Objects

Cited by 26 publications

References 40 publications

Advancements in Steel Surface Defect Detection: An Enhanced YOLOv5 Algorithm with EfficientNet Integration

Advancements in Steel Surface Defect Detection: An Enhanced YOLOv5 Algorithm with EfficientNet Integration

Vision-Based Construction Safety Monitoring Utilizing Temporal Analysis to Reduce False Alarms

Grain rot dataset caused by Burkholderia Glumae Bacteria

Contact Info

Product

Resources

About