期刊
IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS
卷 26, 期 3, 页码 1318-1328出版社
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/JBHI.2021.3104650
关键词
Image segmentation; Deep learning; Lesions; Shape; Generative adversarial networks; Feature extraction; Computer architecture; Deep learning; generative adversarial network; digital holographic imaging; phenotypic analysis of red cells; red cell storage lesions; RBC classification; semantic RBC segmentation; safe transfusions
类别
资金
- National Research Foundation of Korea (NRF) [2020R1A2C3006234]
- Korea Government (MSIT)
- DGIST R&D Program of the Ministry of Science and ICT [21-CoE-BT-02]
- National Research Foundation of Korea [21-COE-BT-02, 2020R1A2C3006234] Funding Source: Korea Institute of Science & Technology Information (KISTI), National Science & Technology Information Service (NTIS)
This study presents a novel approach using digital holographic microscopy to automatically assess the storage lesion of red blood cells (RBCs). The proposed model combines generative adversarial networks (GANs) with marker-controlled watershed segmentation, achieving accurate segmentation and classification of RBCs and complete separation of overlapping cells. The high-throughput method achieved a Dice's coefficient of 0.94 and recognized morphological changes in RBCs during storage.
This study presents a novel approach to automatically perform instant phenotypic assessment of red blood cell (RBC) storage lesion in phase images obtained by digital holographic microscopy. The proposed model combines a generative adversarial network (GAN) with marker-controlled watershed segmentation scheme. The GAN model performed RBC segmentations and classifications to develop ageing markers, and the watershed segmentation was used to completely separate overlapping RBCs. Our approach achieved good segmentation and classification accuracy with a Dice's coefficient of 0.94 at a high throughput rate of about 152 cells per second. These results were compared with other deep neural network architectures. Moreover, our image-based deep learning models recognized the morphological changes that occur in RBCs during storage. Our deep learning-based classification results were in good agreement with previous findings on the changes in RBC markers (dominant shapes) affected by storage duration. We believe that our image-based deep learning models can be useful for automated assessment of RBC quality, storage lesions for safe transfusions, and diagnosis of RBC-related diseases.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据