A time-lapse embryo dataset for morphokinetic parameter prediction

• DOI: 10.1016/j.dib.2022.108258
• Publication Date: 2022-05-11T07:49:29Z
• Authors: Tristan Gomez; Magalie Feyeux; Justine Boulant; Nicolas Normand; Laurent David; Perrine Paul-Gill...; Thomas Fréour; Harold Mouchère
• Keywords: Science (General); Computer applications to medicine. Medical informatics; R858-859.7; Deep learning; 02 engineering and technology; Human reproduction; [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI]; Videos; Q1-390; Time-lapse; IVF; 0202 electrical engineering, electronic engineering, information engineering; Computer vision; Data Article

One of the most common treatments for infertile couples is In Vitro Fertilization (IVF). It consists controlled ovarian hyperstimulation, followed by ovum pickup, fertilization, and embryo culture 2–6 days under environmental conditions, leading to intrauterine transfer or freezing embryos identified as having a good implantation potential embryologists. To allow continuous monitoring development, Time-lapse imaging incubators (TLI) were first released in IVF market around 2010. This time-lapse technology provides dynamic overview embryonic vitro development taking photographs each at regular intervals throughout its development. TLI appears be promising solution improve quality assessment methods, subsequently clinical efficiency IVF. particular, unprecedented high volume high-quality images produced systems has already been leveraged using modern Artificial Intelligence (AI) like deep learning (DL). An important limitation AI-based solutions absence public reference dataset train evaluate (DL) models. this work, we describe fully annotated 704 videos developing with all 7 focal planes available, total 2,4M images. Of note, propose highly detailed annotations 16 different phases, including early cell division but also late divisions, phases after morulation, very which have never used before. that will community morphokinetic models step towards learning-powered We postulate help overall performance DL approaches on ultimately benefiting patients improved success rates.