Denoising Method of Retinal OCT Images Based on Modularized Denoising Autoencoder

Hao Dai; Yaliang Yang; Xian Yue; Shen Chen

doi:10.3788/AOS220815

Acta Optica Sinica, Volume. 43, Issue 1, 0110001(2023)

Denoising Method of Retinal OCT Images Based on Modularized Denoising Autoencoder

Hao Dai^1,2,3, Yaliang Yang^1,2、*, Xian Yue^1,2,3, and Shen Chen^1,2,3

Author Affiliations

¹Institute of Optics and Electronics, Chinese Academy of Sciences, Chengdu 610209, Sichuan , China

²Key Laboratory of Adaptive Optics, Chinese Academy of Sciences, Chengdu 610209, Sichuan , China

³University of Chinese Academy of Sciences, Beijing 100049, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(8)

Fig. 1. Schematic diagram of DAE principle

Download full size

Fig. 2. Network structure of MDAE

Download full size

Fig. 3. Denoised images by using different modules. (a) Original image; (b) T=1; (c) T=2; (d) T=3; (e) T=4

Download full size

Fig. 4. Evaluation metrics varying with number of modules. (a) MSE; (b) PSNR; (c) SSIM

Download full size

Fig. 5. Denoised results of retinal OCT image of normal eye obtained by different methods. (a) Original image; (b) Gaussian filter; (c) mean filter; (d) BM3D; (e) SDAE; (f) MDAE

Download full size

Fig. 6. Denoised results of retinal OCT image of diseased eye obtained by different methods. (a) Original image; (b) Gaussian filter;(c) mean filter; (d) BM3D; (e) SDAE; (f) MDAE

Download full size

Table 1. Quantitative comparison of denoised results obtained by different methods for retinal OCT image of normal eye
View table
Table 1. Quantitative comparison of denoised results obtained by different methods for retinal OCT image of normal eye
Method MSE PSNR /dB SSIM
Original image 861.62 18.78 0.19
Gaussian filter 94.18 28.39 0.59
Mean filter 93.28 28.43 0.58
BM3D 82.54 28.96 0.67
SDAE 82.65 28.95 0.66
MDAE 63.40 30.10 0.69

Table 2. Quantitative comparison of denoised results obtained by different methods for retinal OCT image of diseased eye
View table
Table 2. Quantitative comparison of denoised results obtained by different methods for retinal OCT image of diseased eye
Method MSE PSNR /dB SSIM Time /ms
Original image 963.94 18.29 0.16
Gaussian filter 96.09 28.30 0.55 0.37
Mean filter 91.70 28.51 0.54 1.33
BM3D 72.71 29.54 0.66 6901.00
SDAE 68.97 29.74 0.67 45.71
MDAE 59.72 30.37 0.68 257.12

Tools

Get Citation

Copy Citation Text

Hao Dai, Yaliang Yang, Xian Yue, Shen Chen. Denoising Method of Retinal OCT Images Based on Modularized Denoising Autoencoder[J]. Acta Optica Sinica, 2023, 43(1): 0110001

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Image Processing

Received: Mar. 17, 2022

Accepted: Jun. 20, 2022

Published Online: Jan. 6, 2023

The Author Email: Yang Yaliang (ylyang@ioe.ac.cn)

DOI:10.3788/AOS220815

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology

Table 1. Quantitative comparison of denoised results obtained by different methods for retinal OCT image of normal eye

Table 1. Quantitative comparison of denoised results obtained by different methods for retinal OCT image of normal eye

Table 2. Quantitative comparison of denoised results obtained by different methods for retinal OCT image of diseased eye

Table 2. Quantitative comparison of denoised results obtained by different methods for retinal OCT image of diseased eye