EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

An Image Dataset Quality Evaluation System for Industrial Object Detection Tasks

Shengguo Zhu iD, Yunxi Sun iD, Enhui Lu iD, Xinglong Zhu iD, Jian Liu iD

Sensors · 14.07.2026 · 26 (14) · 4465

DOI: 10.3390/s26144465 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 14.07.2026. Verfasst von Shengguo Zhu und 4 weiteren Autor:innen. Beteiligte Einrichtung: School of Mechanical Engineering, Yangzhou University, Yangzhou 225127, China. Die Arbeit nennt 37 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Planning Project of Science and Technology of Yangzhou, Jiangsu Province Higher Education Basic Science Research General Project.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The performance of visual detection models in industrial applications is strongly influenced by training dataset quality. Although imaging scheme design and algorithm optimization are often emphasized, systematic dataset quality evaluation remains insufficient. To address this gap, this study proposes a dataset quality evaluation framework for industrial object detection. It includes four dimensions and thirteen quantifiable indices: three for acquisition environment, two for image quality, three for dataset scale, and five for annotation quality. Normalization based on theoretical maximum scores is used to reduce biases caused by different score ranges, and dimension weights are assigned using Taguchi orthogonal experiments. Validation is performed on five public and three self-constructed datasets using YOLOv12n and RT-DETR-R18. A positive correlation trend is observed between the proposed scores and detection accuracy, with PLCC/SRCC/Kendall’s tau values of 0.685/0.850/0.764 and 0.656/0.826/0.691, respectively. After second-level weight optimization, the correlations increase to 0.775/0.922/0.837 and 0.748/0.898/0.764. Corresponding p-values and 95% confidence intervals are reported to quantify statistical uncertainty. Sensitivity analysis and ablation comparisons further verify the robustness and necessity of the proposed multidimensional framework. The proposed framework provides a quantifiable method and practical acquisition guidelines for improving industrial image dataset quality.

Schlagwörter

Stichwörter der Autor:innen

Taguchidataset acquisition guidelinesdeep learningimage dataset quality evaluation systemindustrial object detection

Autor:innen und Institutionen

  • School of Mechanical Engineering, Yangzhou University, Yangzhou 225127, China.
  • State Key Laboratory of Advanced Design and Manufacture for Vehicle Body, Hunan University, Changsha 410082, China.

ORCID-Profile: Shengguo Zhu, Yunxi Sun, Enhui Lu, Xinglong Zhu, Jian Liu

Förderung

  • Planning Project of Science and Technology of Yangzhou
  • Jiangsu Province Higher Education Basic Science Research General Project

Zitieren

APA

Zhu, S., Sun, Y., Lu, E., Zhu, X., & Liu, J. (2026). An Image Dataset Quality Evaluation System for Industrial Object Detection Tasks. Sensors, 26(14), 4465. https://doi.org/10.3390/s26144465

Vancouver

Zhu S, Sun Y, Lu E, Zhu X, Liu J. An Image Dataset Quality Evaluation System for Industrial Object Detection Tasks. Sensors. 2026;26(14):4465. doi:10.3390/s26144465

BibTeX
@article{Zhu2026,
  author = {Zhu, Shengguo and Sun, Yunxi and Lu, Enhui and Zhu, Xinglong and Liu, Jian},
  title = {An Image Dataset Quality Evaluation System for Industrial Object Detection Tasks},
  journal = {Sensors},
  year = {2026},
  volume = {26},
  number = {14},
  pages = {4465},
  doi = {10.3390/s26144465}
}
Kostenlos starten →