From Division to Decision: Leveraging Temporal Cell-Stage Segmentation for Embryo Transferability Prediction

2026-05-18

AI Summary

Bovine embryo transfer traditionally relies on a single expert assessment on day 7, resulting in high pregnancy failure rates. To address this limitation, this work proposes TransFACT, a novel framework that, for the first time, incorporates cell developmental stage segmentation as an auxiliary supervision task. Leveraging 2D time-lapse videos from the first four days of development, TransFACT employs a Transformer architecture augmented with temporal convolutions and stage-aware semantic supervision to enable multi-granular temporal modeling. This approach achieves high-accuracy prediction of embryo transferability as early as day 4. Evaluated on a bovine embryo dataset, TransFACT significantly outperforms existing methods, establishing an efficient and reliable new paradigm for early-stage embryo selection.

Abstract

Accurate selection of bovine embryos is a challenging task, as current practice relies on a single expert assessment on the seventh day after insemination, resulting in high rates of pregnancy loss. Time-lapse videomicroscopy provides detailed information on early development, but is difficult to exploit because of complex motion patterns and time-consuming analysis. We propose TransFACT, a transformer-based framework for modeling early developmental stages and embryo transferability using 2D time-lapse videos from the first four days of development. TransFACT combines frame-level temporal features with stage-level representations, using developmental stages as auxiliary supervision to predict transferability on day four. Our experiments demonstrate that TransFACT, by leveraging an existing method designed for action recognition, achieves superior performance than its competitor in predicting embryo transferability.

Problem

• embryo transferability prediction
• bovine embryo selection
• time-lapse videomicroscopy
• pregnancy loss

Innovation

• Transformer-based framework
• temporal cell-stage segmentation
• embryo transferability prediction
• time-lapse videomicroscopy
• auxiliary supervision