A Multi-Strategy Active Learning Framework for Enhanced Peripheral Blood Cell Image Detection

Peripheral blood cell detection is essential for diagnosing and monitoring hematologic disorders. However, existing methods typically focus on a limited number of cell types (usually 3 to 10), restricting their ability to detect a broader range of cells. These methods also struggle with class imbala...

Full description

Saved in:
Bibliographic Details
Main Authors: Yuheng Feng, Jiangtao He, Linjin Wang, Wuchen Yang, Sihan Deng, Lanlin Li, Xinwei Li
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/11036778/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849706058064330752
author Yuheng Feng
Jiangtao He
Linjin Wang
Wuchen Yang
Sihan Deng
Lanlin Li
Xinwei Li
author_facet Yuheng Feng
Jiangtao He
Linjin Wang
Wuchen Yang
Sihan Deng
Lanlin Li
Xinwei Li
author_sort Yuheng Feng
collection DOAJ
description Peripheral blood cell detection is essential for diagnosing and monitoring hematologic disorders. However, existing methods typically focus on a limited number of cell types (usually 3 to 10), restricting their ability to detect a broader range of cells. These methods also struggle with class imbalance and still rely heavily on large amounts of annotated data, limiting their effectiveness for underrepresented categories and hindering scalability. To address these challenges, we propose a novel composite multi-strategy active learning framework using YOLOv5 for enhanced peripheral blood cell detection. The framework reduces annotation costs and improves detection performance by combining uncertainty-based selection, diversity querying, and density-based querying to prioritize the most informative and diverse samples. The process begins with entropy-based uncertainty selection to identify the most uncertain samples, followed by clustering analysis to capture diverse samples from the feature space, and concludes with density-based selection using the k-nearest neighbors algorithm to prioritize samples from high-density regions. Experimental results demonstrate that the framework achieves a mean average precision (mAP@0.5) of 64.8% on a private dataset with 26 cell types, outperforming other active learning strategies and existing methods. It also reduces manual annotation workload by 28.7% compared to random sampling. On the public BCCD dataset (3 cell types), the framework achieves an mAP@0.5 of 86.7%. These results highlight the practicality and reliability of the proposed framework for optimizing peripheral blood cell detection. Our code can be accessed at: <uri>https://github.com/Sar-fyh/bme</uri>
format Article
id doaj-art-c762eca0dc054afeaa2cba0f9c60d0b8
institution DOAJ
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-c762eca0dc054afeaa2cba0f9c60d0b82025-08-20T03:16:18ZengIEEEIEEE Access2169-35362025-01-011310481510482710.1109/ACCESS.2025.357991811036778A Multi-Strategy Active Learning Framework for Enhanced Peripheral Blood Cell Image DetectionYuheng Feng0https://orcid.org/0009-0009-2315-7362Jiangtao He1https://orcid.org/0009-0002-7763-491XLinjin Wang2https://orcid.org/0000-0003-1951-5680Wuchen Yang3Sihan Deng4Lanlin Li5https://orcid.org/0009-0006-3616-3962Xinwei Li6https://orcid.org/0000-0003-0713-9366School of Life Health Information Science and Engineering, Chongqing University of Posts and Telecommunications, Chongqing, ChinaSchool of Life Health Information Science and Engineering, Chongqing University of Posts and Telecommunications, Chongqing, ChinaSchool of Life Health Information Science and Engineering, Chongqing University of Posts and Telecommunications, Chongqing, ChinaMedical Center of Hematology, The Second Affiliated Hospital of Army Medical University, Chongqing, ChinaSchool of Media Arts, Chongqing University of Posts and Telecommunications, Chongqing, ChinaChongqing Defang Information Technology Company Ltd., Chongqing, ChinaSchool of Life Health Information Science and Engineering, Chongqing University of Posts and Telecommunications, Chongqing, ChinaPeripheral blood cell detection is essential for diagnosing and monitoring hematologic disorders. However, existing methods typically focus on a limited number of cell types (usually 3 to 10), restricting their ability to detect a broader range of cells. These methods also struggle with class imbalance and still rely heavily on large amounts of annotated data, limiting their effectiveness for underrepresented categories and hindering scalability. To address these challenges, we propose a novel composite multi-strategy active learning framework using YOLOv5 for enhanced peripheral blood cell detection. The framework reduces annotation costs and improves detection performance by combining uncertainty-based selection, diversity querying, and density-based querying to prioritize the most informative and diverse samples. The process begins with entropy-based uncertainty selection to identify the most uncertain samples, followed by clustering analysis to capture diverse samples from the feature space, and concludes with density-based selection using the k-nearest neighbors algorithm to prioritize samples from high-density regions. Experimental results demonstrate that the framework achieves a mean average precision (mAP@0.5) of 64.8% on a private dataset with 26 cell types, outperforming other active learning strategies and existing methods. It also reduces manual annotation workload by 28.7% compared to random sampling. On the public BCCD dataset (3 cell types), the framework achieves an mAP@0.5 of 86.7%. These results highlight the practicality and reliability of the proposed framework for optimizing peripheral blood cell detection. Our code can be accessed at: <uri>https://github.com/Sar-fyh/bme</uri>https://ieeexplore.ieee.org/document/11036778/Peripheral blood cellsactive learningobject detectiondiversity query
spellingShingle Yuheng Feng
Jiangtao He
Linjin Wang
Wuchen Yang
Sihan Deng
Lanlin Li
Xinwei Li
A Multi-Strategy Active Learning Framework for Enhanced Peripheral Blood Cell Image Detection
IEEE Access
Peripheral blood cells
active learning
object detection
diversity query
title A Multi-Strategy Active Learning Framework for Enhanced Peripheral Blood Cell Image Detection
title_full A Multi-Strategy Active Learning Framework for Enhanced Peripheral Blood Cell Image Detection
title_fullStr A Multi-Strategy Active Learning Framework for Enhanced Peripheral Blood Cell Image Detection
title_full_unstemmed A Multi-Strategy Active Learning Framework for Enhanced Peripheral Blood Cell Image Detection
title_short A Multi-Strategy Active Learning Framework for Enhanced Peripheral Blood Cell Image Detection
title_sort multi strategy active learning framework for enhanced peripheral blood cell image detection
topic Peripheral blood cells
active learning
object detection
diversity query
url https://ieeexplore.ieee.org/document/11036778/
work_keys_str_mv AT yuhengfeng amultistrategyactivelearningframeworkforenhancedperipheralbloodcellimagedetection
AT jiangtaohe amultistrategyactivelearningframeworkforenhancedperipheralbloodcellimagedetection
AT linjinwang amultistrategyactivelearningframeworkforenhancedperipheralbloodcellimagedetection
AT wuchenyang amultistrategyactivelearningframeworkforenhancedperipheralbloodcellimagedetection
AT sihandeng amultistrategyactivelearningframeworkforenhancedperipheralbloodcellimagedetection
AT lanlinli amultistrategyactivelearningframeworkforenhancedperipheralbloodcellimagedetection
AT xinweili amultistrategyactivelearningframeworkforenhancedperipheralbloodcellimagedetection
AT yuhengfeng multistrategyactivelearningframeworkforenhancedperipheralbloodcellimagedetection
AT jiangtaohe multistrategyactivelearningframeworkforenhancedperipheralbloodcellimagedetection
AT linjinwang multistrategyactivelearningframeworkforenhancedperipheralbloodcellimagedetection
AT wuchenyang multistrategyactivelearningframeworkforenhancedperipheralbloodcellimagedetection
AT sihandeng multistrategyactivelearningframeworkforenhancedperipheralbloodcellimagedetection
AT lanlinli multistrategyactivelearningframeworkforenhancedperipheralbloodcellimagedetection
AT xinweili multistrategyactivelearningframeworkforenhancedperipheralbloodcellimagedetection