Real-Time Low-Light Image Enhancement Algorithm for Ophthalmic Surgical Microscope

Tailong Guo; Huaiyu Cai; Yi Wang; Xiaodong Chen

doi:10.3788/AOS241732

Acta Optica Sinica, Volume. 45, Issue 4, 0410001(2025)

Real-Time Low-Light Image Enhancement Algorithm for Ophthalmic Surgical Microscope

Tailong Guo, Huaiyu Cai^*, Yi Wang, and Xiaodong Chen

School of Precision Instrument and Optoelectronics Engineering, Key Laboratory of Optoelectronics Information Technology, Ministry of Education, Tianjin University, Tianjin 300072, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(13)

Fig. 1. Low-illumination image enhancement network architecture. (a) RD-Module architecture; (b) IA-Module architecture; (c) CR-Module architecture

Download full size

Fig. 2. Iterative adjustment curves. (a) n=4; (b) n=8; (c) improved high-order adjustment curves

Download full size

Fig. 3. Cross-attention module

Download full size

Fig. 4. Examples of low light images. (a) Normal light surgical images; (b) corresponding low light synthetic surgical images; (c) normal light pig eye images; (d) corresponding acquired low light pig eye images

Download full size

Fig. 5. Comparative experimental enhancement results of low light synthetic images. (a) Image1; (b) Image2

Download full size

Fig. 6. Comparative experimental enhancement results of low light pig eye images. (a) Image1; (b) Image2

Download full size

Fig. 7. Enhancement results of low light pig eye images under different illumination conditions. (a) 50% reduction in illumination; (b) 70% reduction in illumination; (c) 80% reduction in illumination; (d) 85% reduction in illumination

Download full size

Fig. 8. Experimental results of module ablation. (a) Image1; (b) Image2

Download full size

Fig. 9. Ablation experimental results of loss function. (a) Image1; (b) Image2

Download full size

Table 1. Comparison of evaluation metrics of low light synthetic images

View table

Table 1. Comparison of evaluation metrics of low light synthetic images

Method	HE	MSR	RetinexNet	MBLLEN	KinD	IAT	ZeroDCE	Ours
SSIM	0.4165	0.5828	0.6143	0.8910	0.9121	0.8870	0.7481	0.9123
PSNR /dB	14.0512	8.9725	19.6055	22.3535	23.2392	25.8247	17.4743	29.4408
Params /10⁶	—	—	0.555	0.450	8.160	0.091	0.079	0.571
Time /s	0.0350	11.7000	0.0790	1.1000	1.9800	0.0095	0.0009	0.0070

Table 2. Comparison of evaluation metrics of low light pig eye images
View table
Table 2. Comparison of evaluation metrics of low light pig eye images
Method HE MSR RetinexNet MBLLEN KinD IAT ZeroDCE Ours
SSIM 0.4278 0.6160 0.4357 0.6689 0.6269 0.7337 0.6734 0.7443
PSNR /dB 18.4900 12.8579 13.8123 15.3388 14.9802 21.0790 16.1890 21.7743

Table 3. Comparison of experimental evaluation metrics of module ablation
View table
Table 3. Comparison of experimental evaluation metrics of module ablation
Method Ours Without RD-Module Without CR-Module
SSIM 0.4278 0.6160 0.4357
PSNR /dB 18.4900 12.8579 13.8123

Table 4. Comparison of experimental evaluation metrics of loss function ablation
View table
Table 4. Comparison of experimental evaluation metrics of loss function ablation
Method Ours Without L_col Without L_col1 Without L_per Without L_per1 Without L_iss
SSIM 0.9122 0.7556 0.7705 0.9021 0.8621 0.8999
PSNR /dB 29.4408 27.0037 15.2720 27.4754 27.2749 27.9471

Tools

Get Citation

Copy Citation Text

Tailong Guo, Huaiyu Cai, Yi Wang, Xiaodong Chen. Real-Time Low-Light Image Enhancement Algorithm for Ophthalmic Surgical Microscope[J]. Acta Optica Sinica, 2025, 45(4): 0410001

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites