Nuclear Segmentation in Histopathological Images Using Two-Stage Stacked U-Nets With Attention Mechanism

Abstract

Nuclei segmentation is a fundamental but challenging task in histopathological image analysis. One of the main problems is the existence of overlapping regions which increases the difficulty of independent nuclei separation. In this study, to solve the segmentation of nuclei and overlapping regions, we introduce a nuclei segmentation method based on two-stage learning framework consisting of two connected Stacked U-Nets (SUNets). The proposed SUNets consists of four parallel backbone nets, which are merged by the attention generation model. In the first stage, a Stacked U-Net is utilized to predict pixel-wise segmentation of nuclei. The output binary map together with RGB values of the original images are concatenated as the input of the second stage of SUNets. Due to the sizable imbalance of overlapping and background regions, the first network is trained with cross-entropy loss, while the second network is trained with focal loss. We applied the method on two publicly available datasets and achieved stateof-the-art performance for nuclei segmentation-mean Aggregated Jaccard Index (AJI) results were 0.5965 and 0.6210, and F1 scores were 0.8247 and 0.8060, respectively; our method also segmented the overlapping regions between nuclei, with average AJI = 0.3254. The proposed two-stage learning framework outperforms many current segmentation methods, and the consistent good segmentation performance on images from different organs indicates the generalized adaptability of our approach.

Contextual Mixing Feature Unet for Multi-Organ Nuclei Segmentation

“…And the proposed method achieves 0.8191 on F1 score and 0.608 9 on AJI score. Although the F1 score of our method is lower than RIC-Unet (Zeng et al, 2019) and SUNet (Kong et al, 2020), our AJI score exceeds other methods. Table 3 is comparison results of experiment 2. gLog (Kong et al, 2013) is a unsupervised method and Unet (Ronneberger et al, 2015) and Unet++ (Zhou et al, 2018) are two popular methods used for medical segmentation.…”

“…Unet (Xu et al, 2019) also shares the same encoder and accomplishes the detection task and segmentation task in decoder parts and uses the post-processing step to refine each nucleus. SUNet (Kong et al, 2020) is a two-stage learning framework consisting of two connected, stacked UNets. Pixelwise segmentation masks on nuclei are the first output, and then the output binary masks are combined with original input images to go through the network together again to obtain final results.…”

Hoechst Is All You Need: Lymphocyte Classification with Deep Learning

“…Residual neural networks in general, and the U-Net in particular, are well suited for segmentation tasks as they allow spatial information from the input to propagate directly to the output. Recent work using U-Nets has shown great promise for digital pathology applications [21,20,13,18,1,12], but publications to date have focused on automating tasks that human experts can already do. In this work we take a different approach, asking instead -can a deep neural network learn to do something that human experts cannot?…”

Automatic Semantic Segmentation of the Lumbar Spine: Clinical Applicability in a Multi-parametric and Multi-centre Study on Magnetic Resonance Images

“…Other approaches using ensembles in semantic segmentation tasks are based on transfer learning, where networks trained with different datasets from the one of the target task are retrained (Nigam et al, 2018), or are based on "Stacked U-Nets" trained in two stages. In this last case, ensembles of classifiers are used to detect morphological changes in the cell nucleus from the automatic segmentation of both nuclei regions and regions of overlapping nuclei (Kong et al, 2020). The relevance of ensembles leads to work in which model compression techniques are applied to achieve real-time performance to do predictions in production environments (Holliday et al, 2017).…”