Two-stage Detection Method for Abnormal Cluster Cervical Cells

Abnormal cell detection is a key technique for intelligent assisted diagnosis of cervical cancer, which directly affects the performance of the detection system. However, most cervical abnormal cells exist in the form of clusters. Cells adhere to each other, complex and diverse, which brings challen...

Full description

Saved in:
Bibliographic Details
Main Authors: LIANG Yi-qin, ZHAO Si-qi, WANG Hai-tao, HE Yong-jun
Format: Article
Language:zho
Published: Harbin University of Science and Technology Publications 2022-04-01
Series:Journal of Harbin University of Science and Technology
Subjects:
Online Access:https://hlgxb.hrbust.edu.cn/#/digest?ArticleID=2078
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850052252589359104
author LIANG Yi-qin
ZHAO Si-qi
WANG Hai-tao
HE Yong-jun
author_facet LIANG Yi-qin
ZHAO Si-qi
WANG Hai-tao
HE Yong-jun
author_sort LIANG Yi-qin
collection DOAJ
description Abnormal cell detection is a key technique for intelligent assisted diagnosis of cervical cancer, which directly affects the performance of the detection system. However, most cervical abnormal cells exist in the form of clusters. Cells adhere to each other, complex and diverse, which brings challenges to abnormal cell detection. To solve this problem, we proposed a two-stage detection method for cluster cervical abnormal cells. In the first stage, we use YOLO-v5 target detection network. The standard convolution in the network is replaced by deformable convolution. The size and location of convolution kernel can be dynamically adjusted according to the current pathological image content, so as to adapt to the shape, size and other geometric changes of cervical cells in different clusters. In the second stage, the supervised contrastive learning network is used to learn the feature differences between positive and abnormal clusters of cervical cells, so as to achieve high accuracy classification of positive and abnormal clusters of cervical cells. The experimental results show that the recall rate of cluster cervical cells reaches 89.69 %, which is 1.43 % higher than that of baseline network YOLO-v5.The classification accuracy of positive and abnormal cluster cervical cells reaches 87.81 %, which is 10.31 % higher than that of baseline network ResNet.
format Article
id doaj-art-fa4ecc1181d4429091aa53dd916b70e5
institution DOAJ
issn 1007-2683
language zho
publishDate 2022-04-01
publisher Harbin University of Science and Technology Publications
record_format Article
series Journal of Harbin University of Science and Technology
spelling doaj-art-fa4ecc1181d4429091aa53dd916b70e52025-08-20T02:52:52ZzhoHarbin University of Science and Technology PublicationsJournal of Harbin University of Science and Technology1007-26832022-04-012702768410.15938/j.jhust.2022.02.010Two-stage Detection Method for Abnormal Cluster Cervical CellsLIANG Yi-qin0ZHAO Si-qi1WANG Hai-tao2HE Yong-jun3School of Computer and Technology, Harbin University of Science and Technology, Harbin 150080, ChinaSchool of Computer and Technology, Harbin University of Science and Technology, Harbin 150080, ChinaThe First Affiliated Hospital of Harbin Medical University, Harbin, 150001, ChinaSchool of Computer and Technology, Harbin University of Science and Technology, Harbin 150080, ChinaAbnormal cell detection is a key technique for intelligent assisted diagnosis of cervical cancer, which directly affects the performance of the detection system. However, most cervical abnormal cells exist in the form of clusters. Cells adhere to each other, complex and diverse, which brings challenges to abnormal cell detection. To solve this problem, we proposed a two-stage detection method for cluster cervical abnormal cells. In the first stage, we use YOLO-v5 target detection network. The standard convolution in the network is replaced by deformable convolution. The size and location of convolution kernel can be dynamically adjusted according to the current pathological image content, so as to adapt to the shape, size and other geometric changes of cervical cells in different clusters. In the second stage, the supervised contrastive learning network is used to learn the feature differences between positive and abnormal clusters of cervical cells, so as to achieve high accuracy classification of positive and abnormal clusters of cervical cells. The experimental results show that the recall rate of cluster cervical cells reaches 89.69 %, which is 1.43 % higher than that of baseline network YOLO-v5.The classification accuracy of positive and abnormal cluster cervical cells reaches 87.81 %, which is 10.31 % higher than that of baseline network ResNet.https://hlgxb.hrbust.edu.cn/#/digest?ArticleID=2078cluster cervical cell classificationtarget detectioncontrastive learning
spellingShingle LIANG Yi-qin
ZHAO Si-qi
WANG Hai-tao
HE Yong-jun
Two-stage Detection Method for Abnormal Cluster Cervical Cells
Journal of Harbin University of Science and Technology
cluster cervical cell classification
target detection
contrastive learning
title Two-stage Detection Method for Abnormal Cluster Cervical Cells
title_full Two-stage Detection Method for Abnormal Cluster Cervical Cells
title_fullStr Two-stage Detection Method for Abnormal Cluster Cervical Cells
title_full_unstemmed Two-stage Detection Method for Abnormal Cluster Cervical Cells
title_short Two-stage Detection Method for Abnormal Cluster Cervical Cells
title_sort two stage detection method for abnormal cluster cervical cells
topic cluster cervical cell classification
target detection
contrastive learning
url https://hlgxb.hrbust.edu.cn/#/digest?ArticleID=2078
work_keys_str_mv AT liangyiqin twostagedetectionmethodforabnormalclustercervicalcells
AT zhaosiqi twostagedetectionmethodforabnormalclustercervicalcells
AT wanghaitao twostagedetectionmethodforabnormalclustercervicalcells
AT heyongjun twostagedetectionmethodforabnormalclustercervicalcells