Aim of this research is to apply the deep learning automation based on Time-lapse imaging to jointly assess embryo development，so that it can ensure the consistency of embryo evaluation and improve the accuracy of evaluation.

This study is an observational prospective study after a retrospective analysis. It is a single-center study without randomization or blindness. In the early stage, 1000 patients are collected from three periods of embryo culture through Time-lapse to establish an automated joint evaluation system for the whole process of embryo development. At the later stage, the patients are divided into two groups: Time-Lapse imaging (TLI) +Artificial Intelligence(AI) assessment group and morphological assessment group. 100 patients with Day 5 single blastocyst transplantation are carried out to follow up the pregnancy outcome.

Inclusion Criteria:

• (1) Age < 40 years old; (2) Routine IVF cycles; (3) Period number ≤ 2; (4) The number of ova collected is 5-15; (5) BMI: 18-25 kg/m 2, follicle stimulating hormone(FSH) ≤ 12 IU/L on the third day; (6) Patients with more than 3 high-quality embryos on Day3 and performed single blastocyst transplantation on day 5. (7) Patient without endometrial factors.

Exclusion Criteria:

• (1) Preimplantation Genetic Testing(PGT) is needed due to male infertility, ovulation cycle and chromosome abnormalities; (2) there are systemic diseases of clinical significance; (3) Pictures of blastocysts are not formed or available; (4) Incomplete or unclear image collection in prokaryotic, mitotic and blastocyst phases affected AI evaluation.

Patients are under 40 years old and need artificial intervention to have children