Cell instance segmentation method based on Unet and watershed algorithm

A watershed algorithm and watershed segmentation technology, applied in the field of deep learning

Pending Publication Date: 2021-06-04
HANGZHOU DIANZI UNIV
View PDF0 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The technology of the present invention is aimed at performing instance segmentation on cell images

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Cell instance segmentation method based on Unet and watershed algorithm
  • Cell instance segmentation method based on Unet and watershed algorithm
  • Cell instance segmentation method based on Unet and watershed algorithm

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0023] The present invention will be further described below in conjunction with the accompanying drawings and specific implementation.

[0024] The hardware environment used for implementation is: 8vCPU / 64G memory, the GPU is NVIDIA K80, and the software operating environment is CUDA9, python3.6, pytorch 1.0.

[0025] The data set comes from the BF-C2DL-MUSC of the cell tracking competition

[0026] Such as figure 1 As shown, the first step, the training phase, such as figure 2 , image 3 As shown, a Unet_cell network is trained according to the original image and the cell shape mask to predict the cell shape.

[0027] Such as figure 2 , Figure 4 As shown, a Unet_Nucleus network is trained to predict cell morphology based on the original image and the nucleus mask.

[0028] The second step is to generate a watershed mask. First, the gradient map of the original image is generated, and then the marked map is generated. The marked map points out the water injection poi...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a cell instance segmentation method based on Unet and a watershed algorithm. The cell instance segmentation method is mainly designed for the details of the algorithm combining the Unet with the watershed. According to the method, two unet networks are used and are respectively responsible for semantic segmentation of cellular morphology and semantic segmentation of cell nucleuses; secondly, aiming at the characteristic that a watershed algorithm needing to be marked needs to specify a water injection point and an unknown area; in the algorithm, a cell nucleus is used as a water injection point of each cell region, other cell regions without a cell nucleus region are used as unknown regions, and then an original image is subjected to gradient processing by using a sobel operator to perform watershed algorithm segmentation. The adherent cells can be separated, and the range of the predicted cells is a normal range.

Description

technical field [0001] The invention relates to deep learning, in particular to a cell instance segmentation method based on Unet and watershed. Background technique [0002] Image segmentation algorithms have been developed for many years. From the initial traditional image segmentation algorithm to the current image segmentation algorithm based on deep learning, image segmentation algorithms have made great progress and are widely used in various fields, especially Especially in the field of medical images, segmentation of diseased organs and segmentation and counting of cultured cells have been practically applied. Generally speaking, it is relatively simple to perform deep learning-based semantic segmentation on cell images, and it can be realized with Unet without other operations. However, to perform counting, further counting tasks must be performed, and the cohesive cells in semantic segmentation must be further segmented. Openly, that is, the instance segmentation ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06T7/11G06N3/04G06N3/08G06T5/30
CPCG06N3/04G06N3/08G06T5/30G06T2207/10056G06T2207/20152G06T7/11
Inventor 田泽坤岳雪颖孙玲玲
Owner HANGZHOU DIANZI UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap