Topological Charge Recognition of Vortex Beams Based on Convolutional Neural Network

Tengfei Chai; Xiaoyun Liu; Siyu Gao; Ying Liu; Hongwei Wang; Yumeihui Jin; Yueqiu Jiang

doi:10.3788/AOS250814

Acta Optica Sinica, Volume. 45, Issue 15, 1526001(2025)

Topological Charge Recognition of Vortex Beams Based on Convolutional Neural Network

Tengfei Chai¹, Xiaoyun Liu^1、*, Siyu Gao¹, Ying Liu¹, Hongwei Wang¹, Yumeihui Jin¹, and Yueqiu Jiang^2、**

Author Affiliations

¹School of Science, Shenyang Ligong University, Shenyang 110159, Liaoning , China

²Science and Technology Department, Shenyang Ligong University, Shenyang 110159, Liaoning , China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(9)

Fig. 1. Optical field distributions of vortex beams with various topological charges and spherical aberration coefficients post 2000 m of propagation when $C_{n}^{2} = 10^{- 15} m^{- 2 / 3}$

Download full size

Fig. 2. Structure of SepNet model. (a) Overall structure; (b) architecture of single layer depth separable convolutional module

Download full size

Fig. 3. Evolution of loss value and accuracy with number of iterations. (a) Loss value; (b) accuracy

Download full size

Fig. 4. Three evaluation metrics for $m$ and $k C_{4}$ recognition by SepNet. (a) Different transmission distance; (b) different turbulence intensity

Download full size

Fig. 5. Performance test of SepNet under $C_{n}^{2} = 10^{- 15} m^{- 2 / 3}$

Download full size

Fig. 6. Performance test of SepNet under $C_{n}^{2} = 10^{- 14} m^{- 2 / 3}$

Download full size

Fig. 7. Performance test of SepNet under $C_{n}^{2} = 5 \times 10^{- 15} m^{- 2 / 3}$

Download full size

Table 1. Three evaluation metrics for $m$ ([1.1,5.0]) and $k C_{4}$ recognition across different models

View table

Table 1. Three evaluation metrics for $m$ ([1.1,5.0]) and $k C_{4}$ recognition across different models

Model	Precision		Recall		F1-score
Model	m	kC₄	m	kC₄	m	kC₄
SepNet	100.00%	99.61%	100.00%	99.52%	100.00%	99.56%
Resnet 34	99.96%	98.66%	99.92%	98.29%	99.94%	98.63%
Resnet 18	100.00%	98.67%	100.00%	98.47%	100.00%	98.55%
Xception	100.00%	96.17%	100.00%	96.09%	100.00%	96.07%

Table 2. Three evaluation metrics for $m$ ([1.1,7.0]) and $k C_{4}$ recognition across different models

View table

Table 2. Three evaluation metrics for $m$ ([1.1,7.0]) and $k C_{4}$ recognition across different models

Model	Precision		Recall		F1-score
Model	m	kC₄	m	kC₄	m	kC₄
SepNet	99.95%	99.10%	99.95%	99.06%	99.95%	99.07%
Resnet 34	99.95%	98.56%	99.95%	98.33%	99.95%	98.44%
Resnet 18	99.87%	98.07%	99.84%	97.90%	99.85%	97.97%
Xception	100.00%	95.10%	100.00%	95.22%	100.00%	95.11%

Tools

Get Citation

Copy Citation Text

Tengfei Chai, Xiaoyun Liu, Siyu Gao, Ying Liu, Hongwei Wang, Yumeihui Jin, Yueqiu Jiang. Topological Charge Recognition of Vortex Beams Based on Convolutional Neural Network[J]. Acta Optica Sinica, 2025, 45(15): 1526001

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Physical Optics

Received: Mar. 28, 2025

Accepted: May. 6, 2025

Published Online: Aug. 15, 2025

The Author Email: Xiaoyun Liu (liuxy@imr.ac.cn), Yueqiu Jiang (yueqiujiang@sylu.edu.cn)

DOI:10.3788/AOS250814

CSTR:32393.14.AOS250814

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology

Table 1. Three evaluation metrics for m ([1.1,5.0]) and kC4 recognition across different models

Table 1. Three evaluation metrics for m ([1.1,5.0]) and kC4 recognition across different models

Table 2. Three evaluation metrics for m([1.1,7.0]) and kC4 recognition across different models

Table 2. Three evaluation metrics for m([1.1,7.0]) and kC4 recognition across different models

Table 1. Three evaluation metrics for $m$ ([1.1,5.0]) and $k C_{4}$ recognition across different models

Table 1. Three evaluation metrics for $m$ ([1.1,5.0]) and $k C_{4}$ recognition across different models

Table 2. Three evaluation metrics for $m$ ([1.1,7.0]) and $k C_{4}$ recognition across different models

Table 2. Three evaluation metrics for $m$ ([1.1,7.0]) and $k C_{4}$ recognition across different models