See you in two years!
11-16 April 2026
Conference 13006 > Paper 13006-26
Paper 13006-26

Prediction of abnormal embryonic cell development based on structural similarity coefficient

On demand | Presented live 11 April 2024

Abstract

This study explored the relationship between the structure similarity coefficient extracted from time-lapse monitoring videos of embryos and the embryonic developmental process. It applied the K-means clustering method to analyze the clustering of structure similarity coefficients of different samples, and to investigate the correlation between the clusters with the ability of embryonic cells to develop into blastocysts. The proposed method started to work by using the Hough circle transform to detect cell contours and eliminate image impurities. It further investigated the correlation between the structure similarity coefficient calculated from the time-lapse imaging frames and the embryonic developmental process. In this study, the calculation of the structural similarity coefficient only considered the measure of structural contrast, which accurately reflected the disparity in gray distribution between two images. Normalization was employed to eliminate any influence from brightness and image contrast on the results. After considering the non-uniform distribution of statistical characteristics within an image, local windows were utilized to calculate both mean and variance. We found that the significant decline in the structure similarity coefficient curve corresponds to the event of cell division. The proposed method performed PCA dimensionality reduction on the structure similarity coefficients and applied the K-means clustering to analyze the clustering of sample data. Finally, it explored the relationship between the clustered groups and the ability of embryonic cells to develop into blastocysts. This study generated an effective predictive marker for morphological changes in embryonic cell development, contributing to the prediction of the developmental potential of embryonic cells.

Presenter

Ruipeng Wang
Nankai Univ. (China)
Graduate student, School of Artificial Intelligence, Nankai University
Application tracks: AI/ML
Presenter/Author
Ruipeng Wang
Nankai Univ. (China)
Author
Nankai Univ. (China)
Author
Nankai Univ. (China)
Author
Nankai Univ. (China)
Author
Nankai Univ. (China)
Author
Nankai Univ. (China)