基于YOLOv11模型实现高通量血小板细胞图像精准分割

张思源 ,  马严

中国输血杂志 ›› 2026, Vol. 39 ›› Issue (8) : 1033 -1038.

PDF (1903KB)
中国输血杂志 ›› 2026, Vol. 39 ›› Issue (8) : 1033 -1038. DOI: 10.13303/j.cjbt.issn.1004-549x.2026.08.006
论著

基于YOLOv11模型实现高通量血小板细胞图像精准分割

作者信息 +

Accurate segmentation of high-throughput platelet cell images based on the YOLOv11 model

Author information +
文章历史 +
PDF (1947K)

摘要

目的 建立面向结构光超分辨显微成像高通量数据的血小板自动化实例分割方法,提升宽场明场图像中血小板分割精度。方法 以明场图像为输入,构建基于 YOLOv11 的检测、分割一体化网络,采用统一标注格式进行训练与推理。 网络由主干特征提取、改进多尺度特征融合及检测分割联合预测头组成,以增强小目标与拥挤场景表征。 训练阶段采用预训练权重初始化,输入分辨率 1 024×1 024,训练 200 轮、训练批次为 4 并启用自动混合精度;对灰度 1%与 99%分位数之外像素进行裁剪并线性映射至 8 位;采用翻转、旋转、平移与缩放等增强策略;损失函数为二元交叉熵损失函数与 Dice 损失函数。结果 在胃癌数据集五折交叉验证中,YOLOv11-m 获得较均衡的精确率与召回率,mAP50-95 达到 0. 848;与 Mask R-CNN、YOLOv8-m、YOLOv9-c、YOLO26-m 相比模型权重更小,单图总耗时 14. 85 ms。 模型直接迁移至胆管癌、肝癌、肝硬化及卵巢癌等验证集,检测与分割指标保持稳定,表明该方法在不同数据来源与成像差异下具有泛化能力。结论 本方法可实现血小板高通量、稳定的实例分割,能够有效应对宽场明场血小板显微图像中目标尺度小、密集分布以及局部重叠等复杂场景,并兼顾分割精度与推理效率,可作为血小板图像自动化分析流程中的可靠预处理工具。

Abstract

Objective To develop an automated platelet instance segmentation method for high-throughput structured illumination microscopy (SIM) data, with the goal of improving segmentation accuracy in wide-field bright-field images. Methods Bright-field images were used as input to build a YOLOv11-based unified detection-and-segmentation network, trained and inferred with a standardized annotation format. The network consists of a backbone for feature extraction, an improved PAFPN neck for multi-scale feature fusion, and a joint detection/ segmentation prediction head, enhancing representation of small targets and crowded scenes. Training was initialized with pretrained weights, using an input resolution of 1 024×1 024 for 200 epochs with a batch size of 4 and automatic mixed precision enabled. Pixels outside the 1st and 99th grayscale percentiles were clipped and linearly mapped to 8-bit intensity values. Data augmentation included flipping, rotation, translation, and scaling. The loss function combined binary cross-entropy loss and Dice loss. Results In five-fold cross-validation on the gastric cancer dataset, YOLOv11-m achieved balanced precision and recall, with an mAP50-95 of 0. 848.compared with Mask R-CNN, YOLOv8-m, YOLOv9-c, and YOLO26-m, it used a smaller model size and achieved a total processing time of 14. 85 ms per image. When directly transferred to validation sets of cholangiocarcinoma, hepatocellular carcinoma, liver cirrhosis, and ovarian cancer, the model maintained stable detection and segmentation performance, indicating good generalization across data sources and imaging variations. Conclusion The proposed method enables high-throughput, robust platelet instance segmentation and effectively handles challenging bright-field scenarios involving small targets, dense distributions, and local overlaps. By balancing segmentation accuracy and inference efficiency, it serves as a reliable preprocessing component in automated platelet image-analysis pipelines.

关键词

血小板 / 光学超分辨成像 / 细胞分割 / 深度学习

Key words

platelets / optical super-resolution imaging / cell segmentation / deep learning

引用本文

引用格式 ▾
张思源,马严. 基于YOLOv11模型实现高通量血小板细胞图像精准分割[J]. 中国输血杂志, 2026, 39(8): 1033-1038 DOI:10.13303/j.cjbt.issn.1004-549x.2026.08.006

登录浏览全文

4963

注册一个新账户 忘记密码

参考文献

[1]

WURDINGER T. Tumor-educated platelets[J].Blood, 2019, 133(22):2359-2364.

[2]

BEST M G, WESSELING P, WURDINGER T. Tumor-educated platelets as a noninvasive biomarker source for cancer detection and progression monitoring[J].Cancer Research, 2018, 78(13):3407-3412.

[3]

余恋雨,王宇,唱凯,等. 肿瘤教育血小板源性 miRNA-122 和 miRNA-21 在肝癌早期诊断中的应用研究[J].国际检验医学杂志,2022,43(5):519-522.

[4]

YU L, WANG Y, CHANG K, et al. Application study of tumor-educated platelet-derived miRNA-122 and miRNA-21 in the early diagnosis of hepatocellular carcinoma[J].Inter J Lab Med,2022, 43(5):519-522.

[5]

苟葆城,李萍,马兵. 肿瘤教育血小板联合二代测序技术在肿瘤诊断中的应用[J].医药前沿,2021,11(7):16-17,23.

[6]

GOU B, LI P, MA B. Application of tumor-educated platelets combined with next-generation sequencing in tumor diagnosis[J].Medical Frontier,2021,11(7):16-17,23.

[7]

GLEBOTSKY A C, BLUTEAU D, ESPASANDIN Y R, et al. Mechanisms underlying platelet function defect in a pedigree with familial platelet disorder with a predisposition to acute myelogenous leukemia: potential role for candidate RUNX1 targets[J].J Thrombosis Haemostasis, 2014, 12(5):761-772.

[8]

ZHUGE Y, ZHOU J, YANG G, et al. Activated changes of platelet ultra microstructure and plasma granule membrane protein 140 in patients with non-small cell lung cancer[J].Chin Med J, 2009, 122(9):1026-1031.

[9]

VARGHESE R, SAMBATH M. Yolov8:A novel object detection algorithm with enhanced performance and robustness[C]//2024 International Conference on Advances in Data Engineering and Intelligent Computing systems (ADICS). Chennai, India:IEEE, 2024:1-6.

[10]

LIU Y, Wu J, Chen L. MEIS-YOLO: Improving YOLOv11 for Efficient Aerial Object Detection with Lightweight Design[J].Intelligent and Converged Networks, 2025, 6(2):151-163.

[11]

GAO L, CHEN G. YOLO11-RLN:An aerial UAV algorithm for forest fire detection[J].Annals of the New York Academy of Sciences, 2025,1551(1):312-324.

[12]

MA Y, DENG H, LIU Z, et al. Large-scale quantitative morphometry of platelet α-granules via SIM super-resolution microscopy for cancer liquid biopsy[J].Advanced Science, 2026,13(34).

[13]

XU P, DENG H, Hong Z, et al. Superresolution fluorescence microscopy of platelet subcellular structures as a potential tumor liquid biopsy[J].Small Methods, 2023, 7(10).

[14]

MAO A, MOHRI M, ZHONG Y. Cross-entropy loss functions:Theoretical analysis and applications[C]//International conference on Machine learning. Honolulu Hawaii USA:pmlr, 2023:23803-23828.

[15]

LI X, SUN X, MENG Y, et al. Dice loss for data-imbalanced NLP tasks[C]//Proceedings of the 58th annual meeting of the association for computational linguistics. Online:Association for Computational Linguistics, 2020: 465-476. .

[16]

JOCHER G, QIU J, LIU M, et al. Ultralytics YOLO26: Unified real-time end-to-end vision models[J/OL].(2026-06-02)[2026-07-07]. https://arxiv.org/abs/2606.03748.

[17]

KINGMA D P, BA J. Adam:a method for stochastic optimization[J/OL].(2014-12-22)[2026-07-07]. ht-tps:/ / arxiv.org/abs/ 1412. 6980.

[18]

ZOU Z, CHEN K, SHI Z, et al. Object detection in 20 years:A survey[J].Proceedings of The IEEE, 2023, 111(3):257-276.

[19]

HE, K, GEORGIA G, PIOTR D, et al. Mask r-cnn[C]//Proceedings of The IEEE International Conference on Computer Vision. Venice, Italy:IEEE,2017:2961-2969.

[20]

VARGHESE R, SAMBATH M. YOLOv8:a novel object detection algorithm with enhanced performance and robustness[C]//2024 International Conference on Advances in Data Engineering and Intelligent Computing Systems (ADICS), Chennai, India:IEEE,2024:1-6.

[21]

WANG C Y, YEH I H, LIAO H M. YOLOv9:learning what you want to learn using programmable gradient information[C]//European Conference on Computer Vision. Milan, Italy:Springer, Cham, 2024:1-21.

[22]

SAPKOTA R, CHEPPALLY R H, SHARDA A, et al. YOLO26:key architectural enhancements and performance benchmarking for real-time object detection[J/OL].(2025-09-29)[2026-07-07]. https://arxiv.org/abs/2509.25164.

基金资助

湖北省自然科学基金青年项目(2026AFB461)

AI Summary AI Mindmap
PDF (1903KB)

0

访问

0

被引

详细

导航
相关文章

AI思维导图

/

〈 〉