A Cloud-Based Machine Learning Approach for Blood Cell Classification using YOLOv5

Aug 4, 2026

Accurate blood cell counting is essential for diagnosing health conditions, yet traditional manual microscopy is slow, labor-intensive, and prone to human subjectivity and error. This research presents a cloud-based machine learning system using YOLO (you only look once) for automatic identification and counting of red blood cells, white blood cells, and platelets from blood smear images. A YOLOv5 model is trained on a modified BCCD dataset and evaluated against multiple CNN architectures by measuring accuracy and runtime. Results show rapid counting in under one second and detection performance around 95% precision and 0.99 recall-confidence, supporting practical clinical workflow automation.

LanguageEnglish

CategoryHealthcare

FormatPDF

Pages5

Reading Time13 min

AccessFree to download

Abstract

Introduction

Motivation: limits of manual microscopy

Deep learning solution and YOLOv5 real-time detection

Report objectives and expected impact

Generalizability and data augmentation

A Cloud-Based Machine Learning Approach for Blood Cell Classification using YOLOv5 K. Krishna Jyothi1, G. Kalyani2, A. Sri Karan Chandra3, Akhil Velati4*, C. Srujan5 1,2,3,4,5Dept. of Computer Science and Engineering, Geethanjali College of Engineering and Technology, Hyderabad, India *Corresponding Author: [a](akhil.velati@gmail.com)[khil.velati@gmail.com](akhil.velati@gmail.com), Tel: +91 9347158131 Received: 22/Feb/2024, Accepted: 25/Mar/2024, Published: 30/Apr/2024 Abstract— Checking blood cell counts is crucial for diagnosing health issues. Traditionally, this involves manually counting cells under a microscope, a slow and tiring process. This research explores a new method using machine learning. A machine learning approach for automatic identification and counting of three types of blood cells using ‘you only look once’ (YOLO) object detection and classification algorithm. YOLO framework has been trained with a modified configuration BCCD Dataset of blood smear image to automatically identify and count red blood cells, white blood cells, and platelets. Moreover, this study with other convolutional neural network architectures considering architecture complexity, reported accuracy, and running time with this framework and compare the accuracy of the models for blood cells detection. Overall, the computer-aided system of detection and counting enables us to count blood cells from smear images in less than a second, which is useful for practical applications. Among the state-of-the-arts object detection algorithms such as regions with convolutional neural network (R-CNN), you only look once (YOLO), we chose YOLO framework which is about three times faster than Faster R-CNN with VGG-16 architecture. YOLO uses a single neural network to predict bounding boxes and class probabilities directly from the full image in one evaluation. We retrained YOLO framework to automatically identify and count RBCs, WBCs, and platelets from blood smear images. Also, the trained model has been tested with images from another dataset to observe the precision and accuracy to be around 95% with the recall-confidence to be 0.99. Keywords—RBC; WBC; PLATELETS; CNN; SPPF I. INTRODUCTION Accurate and efficient blood cell analysis plays a critical role in medical diagnosis and patient care. Traditionally, this analysis relies on manual microscopy, a labor-intensive and time-consuming process. Trained technicians visually examine blood smears under a microscope, identifying and counting different types of blood cells like red blood cells (RBCs), white blood cells (WBCs), and platelets. While this method provides valuable information about cell morphology and number, it is inherently subjective and prone to human error. Inconsistencies can arise due to factors like technician fatigue, varying levels of expertise, or difficulties in differentiating certain cell types [2] . Additionally, manual counting is laborious and susceptible to errors, potentially leading to misdiagnosis or delayed treatment. The limitations of manual microscopy have spurred the exploration of alternative approaches. Deep learning offers a compelling solution with the potential to automate blood cell analysis, leading to significant improvements in efficiency and accuracy. This technology allows computers to learn from vast datasets of labeled images, enabling them to identify and classify objects with high precision. By leveraging deep learning models like YOLOv5, we can automate blood cell counting and differentiation, potentially revolutionizing clinical workflows, and expediting patient care [6][9] . This Work delves into the application of YOLOv5, a stateof-the-art deep learning model specifically designed for object detection. YOLOv5, which stands for "You Only Look Once," boasts real-time object detection capabilities. This means it can simultaneously identify and localize objects within an image in a single forward pass through the network. This efficiency

FAQs

为什么需要用机器学习替代传统显微镜手工计数？

传统手工计数耗时费力且受主观因素影响，容易因疲劳、经验差异或细胞类型区分困难而产生错误，从而影响诊断与治疗时效。

该研究使用的YOLOv5在血细胞检测中承担什么任务？

YOLOv5通过单次前向推理直接预测目标边界框与类别概率，用于从血涂片图像中自动识别并计数红细胞、白细胞和血小板。

模型训练与评估的效果如何？

模型在经过训练的BCCD相关数据集上完成检测任务，并在其他数据集上验证泛化；实验报告检测精度与召回表现约为95%精度、0.99召回置信度，同时实现快速计数（小于一秒）。

Aug 4, 2026

Upload by Theodora