AI-Powered Embryo Selection for IVF

Abstract

This project presents a comprehensive AI system for automated embryo analysis and outcome prediction, developed as Lead Computer Vision Algorithm Developer (2023-2024). The system employs state-of-the-art video transformer architectures to analyze time-lapse microscopy videos of embryos, providing objective, data-driven predictions for embryo viability.

The Clinical Challenge

In vitro fertilization (IVF) clinics face a critical decision: which embryo has the highest chance of successful implantation? Traditional embryo assessment relies on manual morphological evaluation by embryologists, which:

Algorithm Portfolio

I developed a comprehensive suite of seven specialized algorithms:

1. End-to-End Video Classification System

Architecture: Video Swin Transformer (3D hierarchical vision transformer)

Tasks: Pregnancy prediction, transfer recommendation, genetic testing outcome

2. Embryo Segmentation Module

Architecture: U-Net encoder-decoder structure

Application: Background removal and normalization

3. Blastocyst Classification

Deep convolutional networks for Day 5-6 embryo quality assessment

4. Comprehensive Embryo Grading System

Multi-criteria quality scoring with clinical integration

5. Morphokinetic Event Detection

Automated detection of developmental milestones using Siamese networks

6. Static Image Prediction Models

Pregnancy outcome prediction from single timepoints

7. Pronuclei Detection System

Early-stage fertilization quality assessment

Video Swin Transformer Architecture

The core innovation is applying Video Swin Transformers to embryo viability prediction—modeling both spatial features and temporal dynamics critical for understanding embryo development.

Key Advantages:

Clinical Impact

The system significantly improves IVF success rates by providing objective, consistent embryo assessment, reducing the number of transfer attempts needed per pregnancy and improving overall clinical outcomes.