Nuclei instance segmentation from histopathology images using Bayesian dropout based deep learning

BMC Medical Imaging

Research

Naga Raju Gudhe¹  Hamid Behravan¹  Arto Mannermaa²  Veli-Matti Kosma² 

[1] Institute of Clinical Medicine, Pathology and Forensic Medicine, Multidisciplinary Cancer research community RC Cancer, University of Eastern Finland, P.O. Box 1627, 70211, Kuopio, Finland;Institute of Clinical Medicine, Pathology and Forensic Medicine, Multidisciplinary Cancer research community RC Cancer, University of Eastern Finland, P.O. Box 1627, 70211, Kuopio, Finland;Biobank of Eastern Finland, Kuopio University Hospital, Kuopio, Finland;

关键词: Semantic segmentation;    Bayesian deep learning;    Uncertainty estimation;    Nuclei segmentation;    Digital pathology;    Medical image analysis;   

DOI  :  10.1186/s12880-023-01121-3

 received in 2023-07-04, accepted in 2023-10-05,  发布年份 2023

来源: Springer

【 摘 要 】

BackgroundThe deterministic deep learning models have achieved state-of-the-art performance in various medical image analysis tasks, including nuclei segmentation from histopathology images. The deterministic models focus on improving the model prediction accuracy without assessing the confidence in the predictions.MethodsWe propose a semantic segmentation model using Bayesian representation to segment nuclei from the histopathology images and to further quantify the epistemic uncertainty. We employ Bayesian approximation with Monte-Carlo (MC) dropout during the inference time to estimate the model’s prediction uncertainty.ResultsWe evaluate the performance of the proposed approach on the PanNuke dataset, which consists of 312 visual fields from 19 organ types. We compare the nuclei segmentation accuracy of our approach with that of a fully convolutional neural network, U-Net, SegNet, and the state-of-the-art Hover-net. We use F1-score and intersection over union (IoU) as the evaluation metrics. The proposed approach achieves a mean F1-score of 0.893 ± 0.008 and an IoU value of 0.868 ± 0.003 on the test set of the PanNuke dataset. These results outperform the Hover-net, which has a mean F1-score of 0.871 ± 0.010 and an IoU value of 0.840 ± 0.032.ConclusionsThe proposed approach, which incorporates Bayesian representation and Monte-Carlo dropout, demonstrates superior performance in segmenting nuclei from histopathology images compared to existing models such as U-Net, SegNet, and Hover-net. By considering the epistemic uncertainty, our model provides a more reliable estimation of the prediction confidence. These findings highlight the potential of Bayesian deep learning for improving medical image analysis tasks and can contribute to the development of more accurate and reliable computer-aided diagnostic systems.

【 授权许可】

CC BY   
© BioMed Central Ltd., part of Springer Nature 2023

【 图 表 】

Fig. 1

Fig. 4

Fig. 2

Fig. 8

12951_2016_246_Article_IEq7.gif

Fig. 1

12951_2016_246_Article_IEq6.gif

Fig. 8

40538_2023_473_Article_IEq1.gif

12951_2015_155_Article_IEq78.gif

Fig. 7

Fig. 3

Fig. 8

Fig. 2

12936_2015_836_Article_IEq25.gif

Fig. 4

12951_2016_246_Article_IEq5.gif

Fig. 2

Fig. 1: The conceptual framework for adherence to treatment guidelines in private drug outlets in Kisumu, Kenya

Fig. 4

12951_2015_155_Article_IEq77.gif

12951_2015_155_Article_IEq76.gif

Fig. 2

Scheme. 1

Fig. 1

Fig. 6

12951_2017_303_Article_IEq1.gif

12951_2016_246_Article_IEq3.gif

Fig. 2

12951_2015_155_Article_IEq74.gif

Fig. 2

Fig. 8

Fig. 5

Fig. 5

Fig. 1

Fig. 2

Fig. 4

Fig. 1

12951_2015_155_Article_IEq72.gif

Fig. 5

12888_2023_5206_Article_IEq3.gif

12888_2023_5206_Article_IEq1.gif

Fig. 9