Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2021 Apr 21;21(9):2921.
doi: 10.3390/s21092921.

Character Recognition of Components Mounted on Printed Circuit Board Using Deep Learning

Sumyung Gang  1 , Ndayishimiye Fabrice  1 , Daewon Chung  2 , Joonjae Lee  3
Affiliations

Character Recognition of Components Mounted on Printed Circuit Board Using Deep Learning

Sumyung Gang et al. Sensors (Basel). .

Abstract

As the size of components mounted on printed circuit boards (PCBs) decreases, defect detection becomes more important. The first step in an inspection involves recognizing and inspecting characters printed on parts attached to the PCB. In addition, since industrial fields that produce PCBs can change very rapidly, the style of the collected data may vary between collection sites and collection periods. Therefore, flexible learning data that can respond to all fields and time periods are needed. In this paper, large amounts of character data on PCB components were obtained and analyzed in depth. In addition, we proposed a method of recognizing characters by constructing a dataset that was robust with various fonts and environmental changes using a large amount of data. Moreover, a coreset capable of evaluating an effective deep learning model and a base set using n-pick sampling capable of responding to a continuously increasing dataset were proposed. Existing original data and the EfficientNet B0 model showed an accuracy of 97.741%. However, the accuracy of our proposed model was increased to 98.274% for the coreset of 8000 images per class. In particular, the accuracy was 98.921% for the base set with only 1900 images per class.

Keywords: PCB inspection; coreset; deep learning; optical character recognition (OCR).

PubMed Disclaimer

Conflict of interest statement

The authors have no conflict of interest relevant to this study to disclose.

Figures

Figure 1
Figure 1
Examples of components attached to PCBs.
Figure 2
Figure 2
The structure of an EfficientNetB0 model with the internal structure of MBConv1 and MBConv6. Compared to MBConv1, MBConv6 has three layers at the top. The number of feature maps as the output is 6.
Figure 3
Figure 3
Examples of images with errors.
Figure 4
Figure 4
Data distribution by collected character class.
Figure 5
Figure 5
Examples of image types in each class.
Figure 6
Figure 6
(a) Data in various colors, (b) Result of color reversal.
Figure 7
Figure 7
V channel distribution: showing the mean number of pixels and standard deviation for classes 0, e, and i.
Figure 8
Figure 8
Random rotation of data at an angle within ±5 degrees.
Figure 9
Figure 9
(a) Example of corrupted data (b) Example of noisy data.
Figure 10
Figure 10
Creation of font data for colors appearing on PCB components in the real world.
Figure 11
Figure 11
Underfitting, overfitting, and optimal fitting.
Figure 12
Figure 12
Data reduction using the grid algorithm. The left part of the figure is a mixture of both original data 1, 2, and augmented data in class 0. The right part of the figure shows the result after applying the n-pick sampling algorithm.
Figure 13
Figure 13
The processes of the proposed method.
Figure 14
Figure 14
Examples of ambiguity between classes.
Figure 15
Figure 15
Visualization of distribution for all classes of the n-Pick (n = 3) dataset and an example of sampled data of class 0.
Figure 16
Figure 16
EfficientNet B7 error cases with 8000 images per class.
See this image and copyright information in PMC

References

    1. Chen X.W., Lin X. Big data deep learning: Challenges and perspectives. IEEE Access. 2014;2:514–525. doi: 10.1109/ACCESS.2014.2325029. - DOI
    1. Zhang Q., Yang L.T., Chen Z., Li P. A survey on deep learning for big data. Inf. Fusion. 2018;42:146–157. doi: 10.1016/j.inffus.2017.10.006. - DOI
    1. Gang S., Fabrice N., Lee J.J. Coresets for PCB Character Recognition based on Deep Learning; Proceedings of the 2020 International Conference on Artificial Intelligence in Information and Communication, ICAIIC 2020; Fukuoka, Japan. 19–21 February 2020; pp. 637–642. - DOI
    1. Fabrice N., Gang S., Lee J.J. Training Data Sets Construction from Large Data Set for PCB Character Recognition. J. Multimed. Inf. Syst. 2019;6:225–234. doi: 10.33851/JMIS.2019.6.4.225. - DOI
    1. Gang S., Lee J. Coreset Construction for Character Recognition of PCB Components Based on Deep Learning. J. Korea Multimed. Soc. 2021;24:382–395.

LinkOut - more resources