Related Experiment Video
Updated: Jul 2, 2026
07:21
Magnetic Resonance Derived Myocardial Strain Assessment Using Feature Tracking
Published on: February 12, 2011
14.3K
A Novel Quadrilateral Contour Disentangled Algorithm for Industrial Instrument Reading Detection
Xiang Li1, Changchang Zeng2, Yong Yao3
1School of Mechanical Engineering, Sichuan University, Chengdu 610065, China.
Entropy (Basel, Switzerland)
|February 26, 2025
Summary
This study introduces the Quadrilateral Contour Disentangled Detection Network (QCDNet) to accurately detect instrument readings in industrial images. QCDNet effectively handles contour distortion and vertex entanglement, improving detection precision and recall.
Area of Science:
- Computer Vision
- Machine Learning
- Industrial Automation
Background:
- Instrument reading detection is challenging due to perspective distortion and vertex entanglement in industrial images.
- Existing methods struggle with accurate automatic display reading because of labeling inaccuracies.
Purpose of the Study:
- To propose a novel network, QCDNet, for robust instrument reading detection.
- To address contour distortion and vertex entanglement issues in industrial instrument images.
Main Methods:
- Developed a Quadrilateral Contour Disentangled Detection Network (QCDNet).
- Utilized a Multi-scale Feature Pyramid Network (MsFPN) for enhanced feature extraction.
- Introduced a Polar Coordinate Decoupling Representation (PCDR) to model contours using polar coordinates and a specialized loss function.
Main Results:
- QCDNet demonstrated superior performance compared to existing quadrilateral detection algorithms.
- Achieved improvements of 4.07% in Precision, 1.8% in Recall, and 2.89% in F-measure on the instrument dataset.
Conclusions:
- QCDNet effectively overcomes challenges in instrument reading detection caused by contour distortion and vertex entanglement.
- The proposed methods, MsFPN and PCDR, contribute to the improved accuracy and robustness of the detection network.
