Multi-Focus Cell Image Fusion Method Based on Target Recognition

Muyan Chen; Jishuai Wang; Lei Wang; Qiang Zhang; Wenchang Xu; Wenbo Cheng

doi:10.3788/LOP242341

Laser & Optoelectronics Progress, Volume. 62, Issue 14, 1417001(2025)

Multi-Focus Cell Image Fusion Method Based on Target Recognition

Muyan Chen^1,2, Jishuai Wang^1,2, Lei Wang^1,2, Qiang Zhang^1,2, Wenchang Xu^1,2, and Wenbo Cheng^1,2、*

¹School of Biomedical Engineering (Suzhou), Division of Life Sciences and Medicine, University of Science and Technology of China, Suzhou 215163, Jiangsu , China

²Suzhou Institute of Biomedical Engineering and Technology, Chinese Academy of Sciences, Suzhou 215163, Jiangsu , China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(12)

Fig. 1. Flow chart of multi-focus cell image fusion algorithm based on target recognition

Download full size

Fig. 2. Improved YOLOv10 structure diagram

Download full size

Fig. 3. AFE module

Download full size

Fig. 4. HFAB

Download full size

Fig. 5. Cell tracking effect diagrams

Download full size

Fig. 6. Image segmentation method

Download full size

Fig. 7. Cell microscopic image sequences. (a) 13 frames of cell microscopic image sequence; (b) 15 frames of cell microscopic image sequence

Download full size

Fig. 8. Comparison of fusion results of different algorithms. (a) Fusion results of 13 frames of cell microscopic image sequence; (b) fusion results of 15 frames of cell microscopic image sequence

Download full size

Table 1. Target recognition results of different models

View table

Table 1. Target recognition results of different models

Model	Accuracy	Precision	Recall	F1-score
SSD	63.39	75.23	80.10	77.59
Fast RCNN	66.53	77.83	82.09	79.90
YOLOv8	86.82	90.95	95.02	92.94
YOLOv9	86.76	91.35	94.53	92.91
YOLOv10	88.48	92.31	95.52	93.89
YOLOv10+HFAB	90.74	92.89	97.51	95.15
YOLOv10+AFE	91.12	93.75	97.01	95.35
YOLOv10+HFAB+AFE	92.96	94.29	98.51	96.35

Table 2. Comparison of evaluation indexes of different algorithms for 13 frames of cell image sequence

View table

Table 2. Comparison of evaluation indexes of different algorithms for 13 frames of cell image sequence

Algorithm	SSIM	PSNR /dB	MSE	RMSE	NRMSE
DCT	0.9877	39.6325	7.0767	2.6602	0.0129
GFF	0.9761	36.0918	15.9919	3.9990	0.0194
LPF	0.9814	21.7037	439.2477	20.9582	0.1121
MWGF	0.9850	38.1901	9.8644	3.1408	0.0152
NSCT	0.9871	39.3746	7.5096	2.7404	0.0133
Fast RCNN	0.9859	40.5936	5.6718	2.3816	0.0115
YOLOv10	0.9875	40.9955	5.1704	2.2739	0.0110
Proposed method	0.9953	45.9693	1.6449	1.2826	0.0062

Table 3. Comparison of evaluation indexes of different algorithms for 15 frames of cell image sequence

View table

Table 3. Comparison of evaluation indexes of different algorithms for 15 frames of cell image sequence

Algorithm	SSIM	PSNR /dB	MSE	RMSE	NRMSE
DCT	0.9525	38.8224	8.5279	2.9203	0.0275
GFF	0.9179	35.1658	19.7923	4.4489	0.0424
LPF	0.9576	27.2358	122.8850	11.0854	0.3167
MWGF	0.9525	37.4913	11.5864	3.4039	0.0314
NSCT	0.9617	38.9102	8.3572	2.8909	0.0705
Fast RCNN	0.9589	36.2699	15.3491	3.9178	0.0189
YOLOv10	0.9618	36.7357	13.7883	3.7133	0.0179
Proposed method	0.9741	41.9349	4.1647	2.0408	0.0144

Table 4. Average running time of different algorithms
View table
Table 4. Average running time of different algorithms
Algorithm Average time Algorithm Average time
DCT 1.6737 NSCT 15.1403
GFF 0.6960 Fast RCNN 0.7843
LPF 0.3011 YOLOv10 0.3007
MWGF 4.5595 Proposed method 0.2459

Tools

Get Citation

Copy Citation Text

Muyan Chen, Jishuai Wang, Lei Wang, Qiang Zhang, Wenchang Xu, Wenbo Cheng. Multi-Focus Cell Image Fusion Method Based on Target Recognition[J]. Laser & Optoelectronics Progress, 2025, 62(14): 1417001

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites