International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 13 Issue: 02 | Feb 2026
p-ISSN: 2395-0072
www.irjet.net
"Diagnosing Blood cells, Classification and Myeloblast Using AI and Deep Learning Method" Prof. Shruthi Rampure1, Bhoomika2 1 2
Professor, Master of Computer Application, VTU, Kalaburagi , Karnataka ,India Student , Master of Computer Application ,VTU, Kalaburagi , Karnataka ,India
-------------------------------------------------------------------------------***----------------------------------------------------------------------style detectors and informs downstream per-cell classifier ABSTRACT- The accurate identification and classification performance under realistic crowding conditions. [4] of blood cells play a crucial role in the early detection and diagnosis of hematological disorders, particularly acute A classic public corpus specifically created for blast-vsmyeloid leukemia (AML). Traditional manual examination of normal discrimination, this dataset is commonly used for peripheral blood smears is time-consuming, labor-intensive, early-stage algorithm benchmarking in leukemia detection and prone to inter-observer variability. To address tasks. Although relatively small, it is well curated and thesechallenges, this study proposes an artificial intelligence provides a focused testbed for methods targeting (AI)-driven framework employing deeplearning techniques for lymphoblast identification. Because of its limited size, careful automated blood cell diagnosis and classification, with a augmentation, crossvalidation, and conservative claims special focus on detecting myeloblasts — a key biomarker for about generalization are necessary. It remains a staple for AML. Convolutional Neural Networks (CNNs), supported by head-to-head comparisons of blast-detection approaches. [3] advanced preprocessing methods such as stain normalization and augmentation, are utilized to classify various blood cell A set of community-curated white-blood-cell collections types, while object detection models such as YOLO are hosted on public platforms, these repositories accelerate integrated to localize and isolate cells from smear images. experimentation by offering varied image styles, class mixes, Transfer learning and class balancing strategies are employed and practical examples for transfer learning. They are to overcome limited blast cell data, improving accuracy and convenient for trying different pretrainedbackbones robustness. The proposed system demonstrates high precision (ResNet, EfficientNet) and for testing preprocessing recipes in distinguishing myeloblasts from other leukocytes and like stain normalization and patch extraction. While generates quantitative outputs such as cell counts and invaluable for development, these collections require proportions, aiding clinical decision-making. harmonization and patient-wise splitting to avoid optimistic evaluation when moving toward clinical deployment. [5] Keyword: Blood Cell Classification, Myeloblast Detection, Acute Myeloid Leukemia (AML), Deep Learning, Artificial Intelligence in Healthcare, Medical Image Analysis
2. PROBLEM STATEMENT Microscopic examination of blood samples is a timeconsuming and labor-intensive process that requires significant effort from skilled hematologists, often leading to diagnostic delays in high-volume clinical environments. Manual interpretation is also highly subjective and prone to human error, as inter-observer variability can cause different experts to classify the same cells differently, while subtle morphological differences in immature cells such as myeloblasts increase the risk of misdiagnosis. Detecting rare cell types presents an additional challenge, since myeloblasts occur infrequently under normal conditions, resulting in class imbalance that can lead to under-detection or incorrect classification, thereby affecting early leukemia diagnosis. Furthermore, many healthcare facilities in rural or resourceconstrained regions lack access to experienced hematologists and advanced diagnostic infrastructure, increasing the likelihood of delayed or missed diagnoses. Although AI-assisted diagnostic systems offer high accuracy, their black-box nature raises concerns regarding transparency and explainability, which may limit clinical trust and slow adoption in critical medical decision-making.
1. INTRODUCTION A large multi-center white-blood-cell collection with multiexpert labels and segmentation masks, this resource provides the volume and annotation quality needed to train robust classifiers and segmentation backbones. Its diversity across imaging devices and annotators helps expose models to real-world variability and reduces overfitting to a single lab. The availability of mask-level labels enables both segmentation and detection workflows, and supports experiments in label-noise mitigation and domain adaptation. It’s a strong choice for developing clinically resilient pipelines. [2] This object-detection oriented dataset provides boundingbox annotations for blood cells, making it ideal for prototyping detection-first pipelines (detect → crop → classify). It’s particularly useful when building two-stage systems that must localize many small objects in dense smears and when evaluating detector throughput and precision. The dataset helps validate yolovX or Faster-RCNN-
© 2025, IRJET
|
Impact Factor value: 8.315
|
ISO 9001:2008 Certified Journal
|
Page 142