Scholarlink

In Vitro Fertilization (IVF) is a complex process that involves retrieving eggs from ovaries and manually combining them with sperm in a laboratory for fertilization. The fertilized eggs, now called embryos, are cultured under strictly controlled conditions for three to five or six days. The highest quality embryo is then transferred into the woman’s uterus. Continuous embryo monitoring with Time-Lapse Imaging (TLI) allows for tracking early embryo development stages, providing valuable time-based development metrics to assess embryo quality before transfer. Experts’ current workflow of manual visual assessment is prone to observer subjectivity and is a time-consuming and challenging process. Integrating high-resolution imaging with Artificial Intelligence (AI) and computer vision techniques opens up new solutions for enhancing IVF’s quality and clinical effectiveness. In this work, we develop a deep learning model to automatically detect the pronuclear and first cleavage stages in IVF embryos, as these stages are critical in the decision-making process. Our experimental results demonstrate the feasibility of automating the detection of these early stages, with the proposed deep learning-based method achieving high accuracy in identifying early embryo development stages.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Deep Neural Networks for Early Human Embryo Stages Detection in Time Lapse Sequences

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Wided Souid Miled,
  • Sana Chtourou,
  • Nozha Chakroun,
  • Khadija Kacem Berjeb

摘要

In Vitro Fertilization (IVF) is a complex process that involves retrieving eggs from ovaries and manually combining them with sperm in a laboratory for fertilization. The fertilized eggs, now called embryos, are cultured under strictly controlled conditions for three to five or six days. The highest quality embryo is then transferred into the woman’s uterus. Continuous embryo monitoring with Time-Lapse Imaging (TLI) allows for tracking early embryo development stages, providing valuable time-based development metrics to assess embryo quality before transfer. Experts’ current workflow of manual visual assessment is prone to observer subjectivity and is a time-consuming and challenging process. Integrating high-resolution imaging with Artificial Intelligence (AI) and computer vision techniques opens up new solutions for enhancing IVF’s quality and clinical effectiveness. In this work, we develop a deep learning model to automatically detect the pronuclear and first cleavage stages in IVF embryos, as these stages are critical in the decision-making process. Our experimental results demonstrate the feasibility of automating the detection of these early stages, with the proposed deep learning-based method achieving high accuracy in identifying early embryo development stages.