Related Experiment Video
Updated: Jan 17, 2026
05:47
Simulation of a Scaled Assembly Process with Collaboration of a Robotic Arm and Monitoring through a Vision System for Quality Control
Published on: August 29, 2025
423
Vision transformer and Mamba-attention fusion for high-precision PCB defect detection.
Asim Niaz1, Muhammad Umraiz1, Shafiullah Soomro2
1Department of Computer Science and Engineering, Chung-Ang University, Seoul, Republic of Korea.
Plos One
|September 25, 2025
Summary
This study introduces ViT-Mamba, a novel computer vision framework for detecting printed circuit board (PCB) defects. It achieves superior accuracy in identifying PCB flaws, enhancing electronic device reliability.
Area of Science:
- Computer Vision
- Artificial Intelligence
- Materials Science
Background:
- Printed circuit board (PCB) defect detection is crucial for consumer electronics reliability.
- Existing deep learning methods face challenges with imbalanced defect data and poor generalization.
- Accurate defect identification requires robust feature extraction and segmentation techniques.
Purpose of the Study:
- To develop an advanced computer vision framework for precise PCB defect detection.
- To overcome limitations of current deep learning models in handling imbalanced defect datasets.
- To enhance the robustness and generalization capabilities of PCB defect identification systems.
Main Methods:
- Proposed ViT-Mamba, a hybrid framework integrating Vision Transformers and Mamba-inspired attention for global feature extraction.
- Introduced an artificial defect generation module to create diverse PCB defect types, improving model robustness.
- Employed a multiscale hierarchical refinement strategy to enhance feature representation for accurate segmentation.
Main Results:
- ViT-Mamba demonstrated superior performance compared to existing methods on a public PCB defect dataset.
- Achieved a high mean Average Precision (mAP) of 99.69%, indicating excellent defect detection accuracy.
- The hybrid approach effectively addressed imbalanced defect distributions and improved generalization.
Conclusions:
- ViT-Mamba offers a powerful and robust solution for automated PCB defect detection.
- The integration of Vision Transformers and Mamba mechanisms significantly advances defect segmentation accuracy.
- This framework holds promise for improving the quality control and reliability of electronic manufacturing.
