Embryo Video Assessment Using AI for Non-Invasive Viability Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current embryo evaluation methods in IVF and animal breeding are subjective, invasive, and lack the ability to accurately assess embryo health and viability, leading to low pregnancy success rates and high costs.
Innovation Solution
A non-invasive, AI/ML-based system that processes real-time video image data of embryos to predict viability, pregnancy likelihood, genetic traits, and offspring characteristics using deep learning models, providing immediate feedback.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional embryo evaluation methods are used, then the process is simple and quick, but the accuracy of assessing embryo health and viability is low and subjective
Solution Approach 1:
The patent replaces traditional mechanical/optical microscopy systems with AI/ML-based deep learning models that process video image data. The system uses neural networks to automatically analyze embryo morphology, movement, and developmental patterns, substituting human subjective evaluation with algorithmic objective assessment, thereby improving measurement precision while managing complexity through software-based solutions.
Solution Approach 2:
The patent introduces video image data as an intermediary between the embryo and the evaluation process. Instead of direct human observation through microscopes, the system captures video footage of embryo movement and morphology, which then serves as input for AI algorithms. This intermediary layer enables automated analysis while preserving the natural state of the embryo, improving assessment accuracy without requiring direct intervention.
2Loss of information
If invasive evaluation methods are used, then more information about embryo quality can be obtained, but the embryo health and viability may be compromised
Solution Approach 1:
The patent replaces invasive mechanical procedures (such as biopsy or direct manipulation) with non-invasive video imaging and AI analysis. The system captures visual and motion data from the embryo's natural behavior in the culture medium, allowing comprehensive quality assessment without physical contact or intervention that could harm the embryo's health or viability.
Solution Approach 2:
The patent creates a digital copy of the embryo's visual and movement characteristics through video imaging. Instead of physically sampling or manipulating the embryo, the system records and analyzes video data that replicates the essential features needed for quality assessment. This digital copying approach provides sufficient information about embryo quality while completely avoiding physical harm.
3Productivity
If multiple embryos are evaluated manually, then detailed assessment can be performed, but the workflow efficiency and time consumption increase
Solution Approach 1:
The patent implements an automated evaluation system where the AI/ML models perform embryo assessment independently without requiring manual intervention for each embryo. The deep learning algorithms automatically process video data, extract relevant features, and generate quality scores, enabling the system to evaluate multiple embryos simultaneously and efficiently, dramatically improving productivity while reducing time loss.
Solution Approach 2:
The patent combines multiple evaluation functions into a single integrated AI system. Instead of separate manual assessments for morphology, movement, and developmental potential, the system merges these evaluation criteria into one unified automated process that analyzes video data comprehensively, improving workflow efficiency without sacrificing assessment detail.
Data Source
AI summary
A method for training a computer-based machine learning model to assess a characteristics-of-interest in embryos such as viability, sex or genetic superiority/inferiority by processing video of each of a plurality of embryos and wherein a subset of the embryos each embodies the characteristic-of-interest. The method includes obtaining (taking or otherwise procuring) a plurality of training videos, each video being of a target embryo and each video having a real-time frame speed (specifically not time-lapse frame speed) over a continuous recording duration of ten minutes or less. Each video includes image data representing micro-movement of the target embryo. The image data is processed thereby generating a computer-based machine learning model for assessing the characteristic-of-interest in future embryos from image data of those embryos.