Cell image segmentation method based on graph path search and deep learning

Abstract

The invention belongs to the technical field of biomedicine and computer image processing, and discloses a cell image segmentation method based on graph path search and deep learning, and the method comprises the following steps: employing a trained U-net prediction model, and carrying out the following steps: in a prediction stage, inputting a to-be-segmented cell image into the trained U-net prediction model, and predicting a distance graph of a to-be-segmented cell; labeling a cell center, and taking the pixel point with the maximum local pixel value as the cell center; searching paths, searching a plurality of paths of two adjacent cell centers, and extracting pixel values of path points; carrying out judging, comparing the pixel value of each path point on the search path with the pixel value of the cell center to judge whether the two cell centers belong to different cells or not, if not, carrying out path search between the other two adjacent cell centers, and if so, carrying out segmentation processing, and repeatedly searching until all the search is finished. According to the invention, the adherent cells in the cell image can be well distinguished and segmented.

Description

technical field [0001] The invention belongs to the technical fields of biomedicine and computer image processing, and in particular relates to a cell image segmentation method based on graph path search and deep learning. Background technique [0002] The automatic segmentation of cell images is of great significance in medical image analysis. Correctly segmenting pathological images of cells can help doctors or researchers identify each cell and study its phenotypic characteristics such as size, color, and shape. Finding out the relationship with characteristics such as genes and diseases will help researchers measure the response of cells to chemicals or in certain biological processes, thereby promoting drug development and shortening the time to market for new drugs. [0003] In recent years, with the continuous in-depth research on deep learning network, its application in cell image segmentation is increasing. There are two main types of deep learning methods for cel...

AI Technical Summary

Problems solved by technology

Since the training cell images often require manual annotation by doctors or medical students, and the workload is large, when the number of cells contained in the cell image is large, it is very difficult to use the instance segmentation network method to segment

In addition, the information of the cell image provided by the instance segmentation network segmentation is less, which is unfavorable for the research of the cell image segmentation

[0008] The existing cell image segmentation methods cannot accurately distinguish cell images with adhesions, and the segmentation effect on cell images is not ideal.

Images