A survey on recent trends in deep learning for nucleus segmentation from histopathology images

Abstract

Nucleus segmentation is an imperative step in the qualitative study of imaging datasets, considered as an intricate task in histopathology image analysis. Segmenting a nucleus is an important part of diagnosing, staging, and grading cancer, but overlapping regions make it hard to separate and tell apart independent nuclei. Deep Learning is swiftly paving its way in the arena of nucleus segmentation, attracting quite a few researchers with its numerous published research articles indicating its efficacy in the field. This paper presents a systematic survey on nucleus segmentation using deep learning in the last five years (2017–2021), highlighting various segmentation models (U-Net, SCPP-Net, Sharp U-Net, and LiverNet) and exploring their similarities, strengths, datasets utilized, and unfolding research areas.

UNETRIS: transformer-based nuclear instance segmentation for three-dimensional fluorescence microscopy images

“…The thresholds we used are {0.25, 0.3, ..., 0.45} for both the Scale 39 -cleared rat kidney and rat liver. These thresholds are specifically chosen as these thresholds are comparable to the thresholds reported in the literature for 3D segmentation 7,21,45,49,[51][52][53][54] . These thresholds are only used for evaluation of the segmentation results and are not used by any of the segmentation methods during training.…”

Automated 3D cytoplasm segmentation in soft X-ray tomography

“…With the increase of AI-based models for automatic segmentation of images from different microscopy techniques, 28 , 33 we compared ACSeg to the increasingly used Segment Anything Model (SAM) developed by Meta AI. 34 As for classical approaches, we selected the same different scenarios of the cells imaged with SXT, see Figure 5 .…”

Improving generalization capability of deep learning-based nuclei instance segmentation by non-deterministic train time and deterministic test time stain normalization

“…The evaluation of histopathological images by experts remains an integral part of the diagnostic routine of many human diseases [15] . An essential element of this process is the inspection of the appearance, morphology, and density of cells, which is subsequently used, for example, to diagnose different types of cancer or to assess the progression of certain diseases [4] , [8] , [34] . Another important aspect in this context is the shape and structure of the cell nuclei [43] .…”

“…Furthermore, considering the recent advances of large language models, vision transformer-based architectures have also been utilized in the encoder-decoder-based models for various histological image analysis tasks, including nuclei instance segmentation [42] , [57] . Comprehensive overviews of state-of-the-art methodologies for nuclei segmentation can be found in the respective studies [4] , [14] , [35] , [57] .…”