Font Size: a A A

Counting Statistics And Morphological Parameters Measurement Technology For Muscle Cell Images

Posted on:2021-05-23Degree:MasterType:Thesis
Country:ChinaCandidate:Full Text:PDF
GTID:2370330647951591Subject:Communication and Information System
Abstract/Summary:PDF Full Text Request
As the world's population gradually grows,mankind's desire for space exploration is becoming stronger.In order to meet the autonomous needs of space missions,explore the effects of space environmental factors on astronaut muscle atrophy and the effect of drug molecules on muscle atrophy.Researchers placed a culture dish containing muscle stem cells suitable for space environments on a satellite,allowing the muscle stem cells to proliferate and divide freely.After the muscle stem cells have proliferated to a certain amount,a new culture medium is replaced and the culture medium contains inducing molecules that can induce the differentiation of muscle stem cells.During cell differentiation,multiple muscle cells fuse to form long tubular myotube.In this process,small drug molecules are added to observe the growth status of myotube,so as to judge the effect of drug molecules on muscle atrophy.In view of this,this paper proposes corresponding algorithms for different needs,including muscle stem cell counting and morphological parameter measurement of myotube.During the proliferation of muscle stem cells,the number of cells needs to be monitored as a basis for changing the culture medium.Traditional algorithms use different filters to preprocess muscle stem cell images,and then use an algorithm based on K-means clustering to count the cells in this paper.However,the traditional method is not good for counting non-circular muscle stem cells,so this paper further proposes a cell detection algorithm based on Mask R-CNN,which uses a variety of data enhancement operators to expand the amount of cell image data and uses filters and binarization algorithms to remove the background and cell information in the image.They enable the data and labels of Usiigaci database to be fused with the cell data in this paper.Finally,an applicable deep learning neural network appears through transfer learning.Compared with traditional methods,this algorithm can identify more types of muscle stem cells,can accurately detect cell edges,and has better data scalability.During the differentiation of muscle cells to form myotube,it's necessary to monitor cell morphology changes to determine the effect of drugs on cell atrophy.An algorithm based on curvature analysis and spline fitting is proposed to automatically measure the morphological parameters of myotube in this paper.Firstly,the cell image is preprocessed to separate the cells from the background,and then use two thresholds to segment the adherent cells to obtain accurate cell boundaries.The cell boundary is fitted by spline interpolation which helped to calculate the normal vector and curvature of the boundary points.Secondly,the matching points based on the normal vector are found to compute the cell width.What's more,the cell endpoints are found by curvature and thus the cytoskeleton is fitted to measure the cell length.The experimental results show that the morphological parameters obtained by the algorithm are consistent with the manual measurement distribution.The algorithm realizes the morphological measurement of myotube,and the average time of cell measurement is shortened by 83.7%,which greatly improves the research and development efficiency of scientific researchers.
Keywords/Search Tags:Muscle Atrophy, Muscle Stem Cells, Myotube, Cell Counting, Morphological Measurements
PDF Full Text Request
Related items