多尺度 CNN-CRF 框架在环境微生物图像分割中的应用。
A Multiscale CNN-CRF Framework for Environmental Microorganism Image Segmentation.
机构信息
Microscopic Image and Medical Image Analysis Group, MBIE College, Northeastern University, Shenyang 110169, China.
Environmental Engineering Department, Northeastern University, Shenyang 110169, China.
出版信息
Biomed Res Int. 2020 Jul 7;2020:4621403. doi: 10.1155/2020/4621403. eCollection 2020.
To assist researchers to identify Environmental Microorganisms (EMs) effectively, a (MSCC) framework for the EM image segmentation is proposed in this paper. There are two parts in this framework: The first is a novel pixel-level segmentation approach, using a newly introduced (CNN), namely, "mU-Net-B3", with a dense (CRF) postprocessing. The second is a VGG-16 based patch-level segmentation method with a novel "buffer" strategy, which further improves the segmentation quality of the details of the EMs. In the experiment, compared with the state-of-the-art methods on 420 EM images, the proposed MSCC method reduces the memory requirement from 355 MB to 103 MB, improves the overall evaluation indexes (Dice, Jaccard, Recall, Accuracy) from 85.24%, 77.42%, 82.27%, and 96.76% to 87.13%, 79.74%, 87.12%, and 96.91%, respectively, and reduces the volume overlap error from 22.58% to 20.26%. Therefore, the MSCC method shows great potential in the EM segmentation field.
为了帮助研究人员有效地识别环境微生物（EMs），本文提出了一种用于 EM 图像分割的 MSCC 框架。该框架有两个部分：第一部分是一种新颖的像素级分割方法，使用新引入的基于卷积神经网络（CNN）的“mU-Net-B3”，并进行密集连接条件随机场（CRF）后处理。第二部分是基于 VGG-16 的补丁级分割方法，采用了一种新颖的“缓冲区”策略，进一步提高了 EM 细节的分割质量。在实验中，与 420 张 EM 图像的最新方法相比，所提出的 MSCC 方法将内存需求从 355MB 降低到 103MB，整体评估指标（Dice、Jaccard、召回率、准确率）从 85.24%、77.42%、82.27%和 96.76%分别提高到 87.13%、79.74%、87.12%和 96.91%，体积重叠误差从 22.58%降低到 20.26%。因此，MSCC 方法在 EM 分割领域具有很大的潜力。
Zotero 插件