Research on Blood Cell Image Detection Method Based on Fourier Ptychographic Microscopy DOI Creative Commons
Mingjing Li, Le Yang, Shu Fang

et al.

Sensors, Journal Year: 2025, Volume and Issue: 25(3), P. 882 - 882

Published: Jan. 31, 2025

Autonomous Fourier Ptychographic Microscopy (FPM) is a technology widely used in the field of pathology. It compatible with high resolution and large field-of-view imaging can observe more image details. Red blood cells play an indispensable role assessing oxygen-carrying capacity human body screening for clinical diagnosis treatment needs. In this paper, cell data set constructed based on FPM system experimental platform. Before training, four enhancement strategies are adopted to improve generalization robustness model. A detection algorithm SCD-YOLOv7 proposed. Firstly, C-MP (Convolutional Max Pooling) module DELAN (Deep Efficient Learning Automotive Network) feature extraction network optimize process ability overlapping features by considering characteristics channels spatial dimensions. Secondly, through Sim-Head head, global information deep map (mean average precision) local details shallow fully utilized performance small target detection. MAP comprehensive indicator evaluating object algorithms, which measures accuracy model calculating precision (AP) under different categories or thresholds. Finally, Focal-EIoU (Focal Extended Intersection over Union) loss function introduced, not only improves convergence speed but also significantly Through quantitative qualitative analysis ablation experiments comparative results, reached 92.4%, increased 7.2%, calculation amount was reduced 14.6 G.

Language: Английский

Research on Blood Cell Image Detection Method Based on Fourier Ptychographic Microscopy DOI Creative Commons
Mingjing Li, Le Yang, Shu Fang

et al.

Sensors, Journal Year: 2025, Volume and Issue: 25(3), P. 882 - 882

Published: Jan. 31, 2025

Autonomous Fourier Ptychographic Microscopy (FPM) is a technology widely used in the field of pathology. It compatible with high resolution and large field-of-view imaging can observe more image details. Red blood cells play an indispensable role assessing oxygen-carrying capacity human body screening for clinical diagnosis treatment needs. In this paper, cell data set constructed based on FPM system experimental platform. Before training, four enhancement strategies are adopted to improve generalization robustness model. A detection algorithm SCD-YOLOv7 proposed. Firstly, C-MP (Convolutional Max Pooling) module DELAN (Deep Efficient Learning Automotive Network) feature extraction network optimize process ability overlapping features by considering characteristics channels spatial dimensions. Secondly, through Sim-Head head, global information deep map (mean average precision) local details shallow fully utilized performance small target detection. MAP comprehensive indicator evaluating object algorithms, which measures accuracy model calculating precision (AP) under different categories or thresholds. Finally, Focal-EIoU (Focal Extended Intersection over Union) loss function introduced, not only improves convergence speed but also significantly Through quantitative qualitative analysis ablation experiments comparative results, reached 92.4%, increased 7.2%, calculation amount was reduced 14.6 G.

Language: Английский

Citations

0