Current applications, limitations, and technical bottlenecks of artificial intelligence in embryo morphological assessment
10.3760/cma.j.cn101441-20241203-00457
- VernacularTitle:人工智能在胚胎形态学评估中的应用现状、局限性与技术瓶颈
- Author:
Mingpeng ZHAO
1
;
Jie LIU
;
Tao ZHANG
Author Information
1. 汉鹏辅助生殖科技有限公司，香港 999077
- Publication Type:Journal Article
- Keywords:
Reproductive technology, assisted;
Embryo selection;
Artificial intelligence;
Blockchain data sharin;
Multimodal integration
- From:
Chinese Journal of Reproduction and Contraception
2025;45(2):135-141
- CountryChina
- Language:Chinese
-
Abstract:
This paper reviews the current applications, technical bottlenecks, and potential breakthrough pathways of artificial intelligence (AI) technology in embryo morphological assessment during in vitro fertilization (IVF). Research indicates that AI technology demonstrates significant advantages in embryo image processing and time-lapse analysis, effectively addressing the subjectivity issues inherent in traditional manual assessment. However, current AI applications still face several challenges, including insufficient data quality and quantity, limited model interpretability, challenges arising from biological diversity, and practical bottlenecks in clinical implementation. To address these limitations, the paper proposes innovative solutions, including blockchain-based global data sharing, data augmentation using generative adversarial networks, multimodal data integration, and virtual laboratory platforms. Through interdisciplinary integration and technological innovation, AI shows promise in achieving personalized and precise embryo selection, potentially improving IVF success rates in the future.
