CN-121999489-A - Feature extraction method and device for trophoblast cells, electronic equipment and storage medium

CN121999489ACN 121999489 ACN121999489 ACN 121999489ACN-121999489-A

Abstract

The application provides a feature extraction method and device of trophoblasts, electronic equipment and a storage medium, and relates to the technical field of assisted reproduction. The method comprises the steps of determining mask images of inner cell clusters, first trophoblast mask images and second trophoblast mask images in target blasts, determining marking information of trophoblast cells in the target blasts according to the first trophoblast mask images, the second trophoblast mask images and the mask images of the inner cell clusters, carrying out curved surface mapping on the mask images of the inner cell clusters, the first trophoblast mask images and the second trophoblast mask images according to the marking information of the trophoblast cells to generate curved surface mapping results, and carrying out feature extraction on trophoblast cell regions in the target blasts according to the curved surface mapping results to obtain multiple three-dimensional features of the trophoblast cell regions. The method can fully excavate more core three-dimensional characteristics of the trophoblast cells, and improves accuracy of blastocyst development potential evaluation.

Inventors

WANG XUEFENG
LIU YAWEN
SHI ZHENZHI

Assignees

广州市华粤行医疗科技有限公司

Dates

Publication Date: 20260508
Application Date: 20260123

Claims (10)

1. A method for extracting characteristics of trophoblasts, comprising: determining a mask image of a cell mass in a target blastula, a first trophoblast mask image and a second trophoblast mask image according to a multifocal target blastula image corresponding to the target blastula; determining marking information of trophoblasts in the target blastocyst according to the first trophoblast mask image, the second trophoblast mask image and the mask image of the inner cell mass; performing surface mapping on the mask image of the inner cell mass, the mask image of the first trophoblast cell and the mask image of the second trophoblast cell according to the marking information of the trophoblast cell to generate a surface mapping result, wherein the surface mapping result is used for indicating the region of the inner cell mass and the trophoblast cell, which is located on a surface model of a target blastula after being mapped to the surface model; And extracting characteristics of the trophoblast cell area in the target blastula according to the curved surface mapping result to obtain a plurality of three-dimensional characteristics of the trophoblast cell area.
2. The method of claim 1, wherein determining the mask image of the inner cell mass, the first trophoblast mask image, and the second trophoblast mask image in the target blastocyst from the multifocal target blastocyst image corresponding to the target blastocyst comprises: Performing image fusion processing on the multi-focus target blastula image corresponding to the target blastula to obtain a first fusion image and a second fusion image; performing image definition analysis on each coke target blastula image, determining a target coke blastula image, and dividing to obtain a mask image of the inner cell mass based on the target coke blastula image; And respectively carrying out trophoblast cell segmentation on the first fusion image and the second fusion image to obtain a first trophoblast cell mask image and a second trophoblast cell mask image.
3. The method of claim 2, wherein performing an image sharpness analysis on each of the focus target blastula images to determine a target focus blastula image comprises: Dividing each focus target blastula image into a plurality of sub-image blocks, and determining the image definition of each sub-image block corresponding to each focus target blastula image; determining the image definition of each focus target blastula image according to the image definition of each sub-image block corresponding to each focus target blastula image; and determining the target coke blastula image from each coke target blastula image according to the image definition of each coke target blastula image.
4. The method of claim 1, wherein determining the marker information for trophoblasts in the target blastocyst based on the first trophoblast mask image, the second trophoblast mask image, and the mask image for the inner cell mass comprises: Determining marking information of each trophoblast in the first trophoblast mask image according to the first trophoblast mask image and the mask image of the inner cell mass, wherein the marking information is used for indicating that the trophoblast is an extremely trophoblast or a parietal trophoblast; and determining the marking information of each trophoblast in the second trophoblast mask image according to the second trophoblast mask image and the mask image of the inner cell mass.
5. The method of claim 1, wherein the performing surface mapping on the mask image of the inner cell mass, the first trophoblast mask image, and the second trophoblast mask image based on the marking information of the trophoblasts, and generating a surface mapping result comprises: According to the coordinates of each pixel point in the mask image of the inner cell mass, determining coordinate information of each pixel point in the mask image of the inner cell mass after mapping to the curved surface model of the target blastula; determining each pixel point corresponding to each polar trophoblast cell and each pixel point corresponding to each wall trophoblast cell in the first trophoblast mask image according to the marking information of the trophoblast cells; determining each pixel point corresponding to each polar trophoblast and each pixel point corresponding to each wall trophoblast in the second trophoblast mask image; According to each pixel point corresponding to each polar trophoblast cell and each pixel point corresponding to each wall trophoblast cell in the first trophoblast cell mask image, determining coordinate information after each pixel point corresponding to each polar trophoblast cell in the first trophoblast cell mask image is mapped to a curved surface model of the target blastula and coordinate information after each pixel point corresponding to each wall trophoblast cell is mapped to the curved surface model of the target blastula; Determining coordinate information of each pixel point corresponding to each polar trophoblast cell in the second trophoblast mask image mapped to the curved surface model of the target blastula and coordinate information of each pixel point corresponding to each wall trophoblast cell mapped to the curved surface model of the target blastula according to each pixel point corresponding to each polar trophoblast cell and each pixel point corresponding to each wall trophoblast cell in the second trophoblast mask image; And generating the curved surface mapping result according to the coordinate information of each pixel point in the mask image of the inner cell mass after mapping to the curved surface model of the target blastula, the coordinate information of each pixel point in the first trophoblast mask image after mapping to the curved surface model of the target blastula, the coordinate information of each pixel point in the second trophoblast mask image after mapping to the curved surface model of the target blastula, and the coordinate information of each pixel point in the second trophoblast mask image after mapping to the curved surface model of the target blastula.
6. The method according to claim 5, wherein determining the coordinate information of each pixel in the mask image of the inner cell mass mapped to the curved surface model of the target blastocyst according to the coordinates of each pixel in the mask image of the inner cell mass comprises: Determining a first angle and a second angle according to coordinates and scaling factors of target pixel points in the mask image of the inner cell mass, wherein the first angle is used for determining the position of the target pixel points in the horizontal direction when the target pixel points are mapped onto the curved surface model of the target blastula, and the second angle is used for determining the height information of the target pixel points in the vertical direction when the target pixel points are mapped onto the curved surface model of the target blastula; and determining coordinate information of the target pixel point mapped to the curved surface model of the target blastula according to the first angle and the second angle.
7. The method according to claim 1, wherein the feature extraction of the trophoblast cell region in the target blastocyst according to the curved mapping result, to obtain a plurality of three-dimensional features of the trophoblast cell region, comprises: Determining corresponding curved surface areas of all the polar trophoblasts in the curved surface model of the target blastula and corresponding curved surface areas of all the wall trophoblasts in the curved surface model of the target blastula according to the curved surface mapping result; determining at least one characteristic of the polar trophoblast cell region, namely the area of the polar trophoblast cell region, the perimeter of the polar trophoblast cell region, the roundness of the polar trophoblast cell region, the width of the polar trophoblast cell region, the length of the polar trophoblast cell region and the ratio of the short axis to the long axis of the polar trophoblast cell region according to the corresponding curved surface region of each polar trophoblast cell in the curved surface model of the target blastula; Determining at least one characteristic of the wall trophoblast cell region, namely the area of the wall trophoblast cell region, the perimeter of the wall trophoblast cell region, the roundness of the wall trophoblast cell region, the width of the wall trophoblast cell region, the length of the wall trophoblast cell region and the ratio of the short axis to the long axis of the wall trophoblast cell region according to the corresponding curved surface region of each wall trophoblast cell in the curved surface model of the target blastula; And determining a plurality of three-dimensional characteristics of the trophoblast cell region according to the characteristics of the polar trophoblast cell region and the characteristics of the wall trophoblast cell region.
8. The characteristic extraction device of the trophoblast is characterized by comprising a determination module, a generation module and a characteristic extraction module; the determining module is used for determining a mask image of an inner cell mass in the target blastula, a first trophoblast mask image and a second trophoblast mask image according to a multi-focal target blastula image corresponding to the target blastula; the determining module is used for determining the marking information of the trophoblasts in the target blastocyst according to the first trophoblast mask image, the second trophoblast mask image and the mask image of the inner cell mass; The generating module is used for carrying out curved surface mapping on the mask image of the inner cell mass, the mask image of the first trophoblast cell and the mask image of the second trophoblast cell according to the marking information of the trophoblast cell to generate a curved surface mapping result, wherein the curved surface mapping result is used for indicating the area of the inner cell mass and the trophoblast cell, which is located on a curved surface model after being mapped to the curved surface model of a target blastula; and the characteristic extraction module is used for extracting characteristics of the trophoblast cell area in the target blastula according to the curved surface mapping result to obtain a plurality of three-dimensional characteristics of the trophoblast cell area.
9. An electronic device comprising a processor, a storage medium and a bus, wherein the storage medium stores program instructions executable by the processor, the processor and the storage medium communicating via the bus when the electronic device is in operation, the processor executing the program instructions to implement the method of extracting characteristics of trophoblast cells according to any one of claims 1 to 7.
10. A computer readable storage medium, characterized in that the storage medium has stored thereon a computer program which, when executed by a processor, is adapted to carry out the method of feature extraction of trophoblasts according to any one of claims 1 to 7.

Description

Feature extraction method and device for trophoblast cells, electronic equipment and storage medium Technical Field The application relates to the technical field of assisted reproduction, in particular to a method and a device for extracting characteristics of trophoblast cells, electronic equipment and a storage medium. Background In assisted reproduction technology, embryo development potential evaluation is a key link for determining success or failure of embryo implantation, and trophoblast cells in blastula are important factors for determining success or failure of blastula implantation, and especially morphological characteristics of extremely trophoblast cells are closely related to development and implantation potential of blastula. In the traditional method, the evaluation of embryo development potential is realized based on morphological evaluation, and the indexes of the embryo are mainly including the expansion degree of the blastula cavity, the morphological classification of inner cell mass and trophoblast cells, and the subjective observation of blastula on a two-dimensional plane by embryologists. The method has strong subjectivity, and the characteristic information of the trophoblast cells is extracted only to a limited extent, so that the accuracy of the evaluation result of embryo development potential is low. Disclosure of Invention The application aims at overcoming the defects in the prior art, and provides a method, a device, electronic equipment and a storage medium for extracting characteristics of trophoblast cells so as to extract core three-dimensional characteristics of the trophoblast cells and improve the evaluation accuracy of embryo development potential. In order to achieve the above purpose, the technical scheme adopted by the embodiment of the application is as follows: in a first aspect, an embodiment of the present application provides a method for extracting characteristics of a trophoblast cell, including: determining a mask image of a cell mass in a target blastula, a first trophoblast mask image and a second trophoblast mask image according to a multifocal target blastula image corresponding to the target blastula; determining marking information of trophoblasts in the target blastocyst according to the first trophoblast mask image, the second trophoblast mask image and the mask image of the inner cell mass; performing surface mapping on the mask image of the inner cell mass, the mask image of the first trophoblast cell and the mask image of the second trophoblast cell according to the marking information of the trophoblast cell to generate a surface mapping result, wherein the surface mapping result is used for indicating the region of the inner cell mass and the trophoblast cell, which is located on a surface model of a target blastula after being mapped to the surface model; And extracting characteristics of the trophoblast cell area in the target blastula according to the curved surface mapping result to obtain a plurality of three-dimensional characteristics of the trophoblast cell area. Optionally, the determining the mask image of the inner cell mass, the first trophoblast mask image and the second trophoblast mask image in the target blastocyst according to the multi-focal target blastocyst image corresponding to the target blastocyst includes: Performing image fusion processing on the multi-focus target blastula image corresponding to the target blastula to obtain a first fusion image and a second fusion image; performing image definition analysis on each coke target blastula image, determining a target coke blastula image, and dividing to obtain a mask image of the inner cell mass based on the target coke blastula image; And respectively carrying out trophoblast cell segmentation on the first fusion image and the second fusion image to obtain a first trophoblast cell mask image and a second trophoblast cell mask image. Optionally, the image fusion processing is performed on the multifocal target blastula image corresponding to the target blastula to obtain a first fusion image and a second fusion image, including: Collecting multi-focus original blastula images corresponding to target blastula, and performing blastula main body segmentation on each focus original blastula image to obtain each focus blastula image; performing image enhancement processing on each coke blastula image to obtain each coke target blastula image; grouping the target blastula images of each focus according to a focus plane division strategy to obtain a first group of images and a second group of images; And carrying out fusion processing on each focus target blastula image in the first group of images to obtain the first fusion image, and carrying out fusion processing on each focus target blastula image in the second group of images to obtain the second fusion image. Optionally, the image sharpness analysis is performed on each focal target blastula image to determine a target