Systems and methods for generating a mask for automated assessment of embryo quality

Patent metadata

Field	Value
Publication number	US-10510143-B1
Application number	US-201615271054-A
Country	US
Kind code	B1
Filing date	Sep 20, 2016
Priority date	Sep 21, 2015
Publication date	Dec 17, 2019
Grant date	Dec 17, 2019

Abstract

Systems and methods for generating a mask for automated assessment of embryo quality are disclosed herein. The method for generating a mask for automated assessment of embryo quality can include receiving an image, including a plurality of pixels, of a human embryo from an imaging system. A pixel can be selected and features of the selected pixel can be determined by generating a plurality of random boxes of random sizes and at random locations about the selected pixel. The selected pixel can be identified as one of: inside of a mask area; and outside of the mask area based on the determined features.

First claim

What is claimed is:

1. A method for determining viability of a human embryo, the method comprising:
receiving an image comprising a human embryo from an imaging system, said image comprising a plurality of pixels, wherein at least one pixel is selected from said plurality of pixels;
generating a plurality of random boxes on said image, said plurality of random boxes are located around said at least one selected pixel, and wherein said plurality of random boxes have a plurality of random sizes and a plurality of random locations;
identifying that said at least one selected pixel within at least one of said plurality of random boxes is inside said at least one human embryo or outside said at least one human embryo such that an embryo mask is superimposed on the received image wherein the embryo mask distinguishes between a first portion of the image and a second portion of the image, wherein the first portion contains the image of the embryo, and wherein the second portion does not contain the image of the embryo;
pairing said plurality of random boxes to detect a feature of the image based on the first portion of the image; and
generating a viability prediction based on the detected feature of the image such that the viability prediction recommends selection of the human embryo based upon a predicted likelihood of implantation; and
implanting the selected human embryo in a human.

2. The method of claim 1 , wherein the image based feature is selected from the group consisting of embryo image area; cavity image area; a change in embryo image area over time; a change in cavity image area over time; embryo image perimeter; and convex hull.

3. The method of claim 1 , wherein the image based feature is selected from the group consisting of cavitation; hatching; embryo expansion; and embryo collapse.

4. The method of claim 1 , wherein the image-based features is selected from the group consisting of an area of the embryo; an area of a cavity of the embryo; a perimeter of the embryo; and a convex hull.

5. The method of claim 1 , wherein the viability prediction further comprises a prediction of euploidy in the human embryo.

6. An imaging system for evaluation of a human embryo, the system comprising:
a stage configured to receive a multi-well culture dish comprising a plurality of micro-wells, wherein each of said plurality of micro-wells contain at least one human embryo;
a time-lapse microscope configured to:
acquire a series of time-lapse images of the at least one human embryo;
select an image of the at least one human embryo from the series of time-lapse images, said image comprising a plurality of pixels;
generate a plurality of random boxes having a random location and a random size that are located around said plurality of pixels to superimpose an embryo mask on the received image, wherein the embryo mask distinguishes between a first portion of the image and a second portion of the image;
detect a feature of the image based on at least one pairing of said plurality of random boxes; and
generate a viability prediction that determines a developmental potential for implantation of the at least one human embryo based on the detected image based feature; and

a mask classifier software module configured to pair said plurality of random boxes around at least one pixel of said plurality of pixels to detect at least one feature of said at least one embryo.

7. The imaging system of claim 6 , wherein the image based feature is selected from the group consisting of embryo image area; cavity image area; a change in embryo image area over time; a change in cavity image area over time; embryo image perimeter; and convex hull.

8. The imaging system of claim 6 , wherein the image based feature is selected from the group consisting of cavitation; hatching; embryo expansion; and embryo collapse.

9. The imaging system of claim 6 , wherein the image based feature is selected from the group consisting of an area of the embryo; an area of a cavity of the embryo; a perimeter of the embryo; and a convex hull.

Assignees

• Progyny Inc
• Ares Trading Sa

Inventors

• Zhou Yan
• Chian Martin T
• Tan Lei
• Koppel Daniel E

Classifications

• G06T2207/30044

  Fetus; Embryo · CPC title

• G06T7/0012 Primary

  Biomedical image inspection · CPC title

• G01N33/4833 Primary

  of solid biological material, e.g. tissue samples, cell cultures (tissue in vivo A61B5/00; cell suspensions G01N33/48735) · CPC title

• G06T2207/10056

  Microscopic image · CPC title

• G06T7/60

  Analysis of geometric attributes · CPC title