HUSCAP logo Hokkaido Univ. logo

Hokkaido University Collection of Scholarly and Academic Papers >
Graduate School of Health Sciences / Faculty of Health Sciences >
Peer-reviewed Journal Articles, etc >

Deep Learning-Based Nuclear Lobe Count Method for Differential Count of Neutrophils

Files in This Item:

The file(s) associated with this item can be obtained from the following URL:

Title: Deep Learning-Based Nuclear Lobe Count Method for Differential Count of Neutrophils
Authors: Yabuta, Mayu Browse this author
Nakamura, Iori Browse this author
Ida, Haruhi Browse this author
Masauzi, Hiromi Browse this author
Okada, Kazunori Browse this author
Kaga, Sanae Browse this author →KAKEN DB
Miwa, Keiko Browse this author →KAKEN DB
Masauzi, Nobuo Browse this author
Keywords: blood cell automatic image analysis
computer vision
convolutional neural networks
deep learning
white blood cell morphology
Issue Date: Jul-2021
Publisher: Tohoku University Medical Press
Journal Title: The Tohoku journal of experimental medicine
Volume: 254
Issue: 3
Start Page: 199
End Page: 206
Publisher DOI: 10.1620/tjem.254.199
Abstract: Differentiating neutrophils based on the count of nuclear lobulation is useful for diagnosing various hematological disorders, including megaloblastic anemia, myelodysplastic syndrome, and sepsis. It has been reported that one-fifth of sepsis-infected patients worldwide died between 1990 and 2017. Notably, fewer nuclear-lobed and stab-formed neutrophils develop in the peripheral blood during sepsis. This abnormality can serve as an early diagnostic criterion. However, testing this feature is a complex and time-consuming task that is rife with human error. For this reason, we apply deep learning to automatically differentiate neutrophil and nuclear lobulation counts and report the world's first small-scale pilot. Blood films are prepared using venous peripheral blood taken from four healthy volunteers and are stained with May Grunwald Giemsa stain. Six-hundred 360 x 363-pixel images of neutrophils having five different nuclear lobulations are automatically captured by Cellavision DM-96, an automatic digital microscope camera. Images are input to an original architecture with five convolutional layers built on a deep learning neural-network platform by Sony, Neural Network Console. The deep learning system distinguishes the four groups (i.e., band-formed, two-, three-, and four- and five-segmented) of neutrophils with up to 99% accuracy, suggesting that neutrophils can be automatically differentiated based on their count of segmented nuclei using deep learning.
Type: article
Appears in Collections:保健科学院・保健科学研究院 (Graduate School of Health Sciences / Faculty of Health Sciences) > 雑誌発表論文等 (Peer-reviewed Journal Articles, etc)

Export metadata:

OAI-PMH ( junii2 , jpcoar_1.0 )

MathJax is now OFF:


 - Hokkaido University