Home > > > AI in Embryo Assessment

AI in Embryo Assessment

Author : Medline Academics | Published On : 20 Jul 2026

It's a great pleasure to be with you today and I feel very privileged to be part of this international conference. Today I am talking about artificial intelligence and embryo assessment and really when I started to search and prepare this lecture, I found the topic is very deep and really, I feel that I need to edit a book, not only a lecture, because it's a very promising subject.

Many students and professionals interested in ART often ask about Embryology Course Fees before choosing a training program. While IVF Course fees vary depending on the institution, curriculum, clinical exposure, and certification, selecting a comprehensive embryology course with hands-on laboratory training is more important than focusing on cost alone. Understanding advanced topics such as AI in embryo assessment is becoming an essential part of modern embryology education.

Artificial intelligence has been implemented in all the medical specialities and resulted in significant impact in the practise of medicine. This includes the assisted reproduction technology. In assisted reproduction technology, artificial intelligence is used to predict the success of ART cycles. That is at the start of the cycle when a patient come to see the physicians and depending on the patient history and the status, it can predict the outcome of the cycle and it can predict the oocyte cryosurgical and development competence. It can judge on every single oocyte, depends on the patient history and the oocyte morphology. And it helps in embryo assessment, which is the focus of our talk today. Embryo assessment includes the morphology, morphometry, morpho kinetics, and the chromosomal contents, the PGTA. And it helps also in lab performance and the KPIs. It can help to make KPIs for the physicians and for the embryologists, as well as in quality control.

Artificial intelligence is defined as a machine that uses multiple tools, including machine learning, to think rationally and act rationally like a human. So, the target of artificial intelligence is to produce machines that can think like us as embryologists and can take decisions or make suggestions, but in a faster way.

To develop artificial intelligence, the first step is to have scientific data. Scientific data will lead to deep learning and the deep learning will enable us to feed the machine with the deep knowledge we have and to give the machine learning and finally will give artificial intelligence.

So, the scientific data. At the start when we used to transfer embryos at the blastocyst stage, at the very early stages, we used to transfer all types of blastocysts. So, at that time we didn't wish blastocysts will give higher, will give a pregnancy and have much chances to give a viable live birth. After having a lot of data and pictures and analysing them, we have deep learning. In this deep learning we know that such blastocysts in the picture have lower chance of giving a live birth or causing pregnancy and these two blasts on the right-hand side have higher chance.

We did this after making retrospective analysis of the blastocysts that were transferred to patients. So and after that we will go to the machine learning. Machine learning is considered as a sophisticated statistical model that used for a prediction and it keeps improving by time. That is to say the more we work, the more we give information to machine, the more we have better results and better predictions. So, machine learning has three main levels.

  • So, the level one is called the unsupervised learning. In this level, the system is given examples and know how to differentiate between the general knowledge. In the embryology, we teach the machine how to differentiate the oocyte and various type of oocytes and embryos and the blastocysts. So, this is general knowledge at level one of the machine learning.
  • The second level is supervised learning. We give the system examples and we show the system the output of the examples. Here we have one good blastocyst.

We give it to like several pictures to the machine and we give the result to the machine that this quality of a blastocyst can give us a healthy live bear. And the other thing is this quality of a blastocyst will give us a negative HCG test. So, we teach the machine and after the machine system is taught properly and taking thousands of pictures and information from us, it will be able to give us a prediction.

  • And the third level is the reinforced learning. This system takes action and learns from the good or bad outcome. As we give the machine early in the previous slide, we showed it the good slide, the good blast and the bad blast.

This will enable the machine to give a prediction and the scores of the possibility of having pregnancy. The machine gives us a probability for pregnancy for each individual embryo. The lowest has a chance of 20% and the highest has a chance of around 90%. This is according to the prediction of the machine. And the machine learning will give us the artificial intelligence embryo selection. The machine learning tool led to the development of software that can rank best embryos for each patient.

And there are a number of commercially available software developed by time-lapse manufacturer and even by IT suppliers. I had the chance to talk with time-lapse manufacturer, the three main ones in the market. And all of them have the software that can predict and help for embryo selection.

The software is objective. This means they judge on every embryo on the same way. And if you give the same embryo after several times, it will give the same result.

But nothing is 100% guaranteed. Like the system cannot guarantee it's 100% you will get the pregnancy or live birth. So, after we check the embryo morphology, and the embryo is beautiful, it is not enough that the embryo will implant.

We have a lot of beautiful embryos and good-looking embryos that are unemployed. So, the artificial intelligence helps in improving the results of the next-generation sequencing result for PGTA. Usually, in most of the results of the PGTA, the results can be likely, but in some embryos, an interpretation varies among the geneticists. The collection of data and artificial intelligence help us to improve the interpretation errors in the next-generation sequencing. And it limits the subjective interpretation, especially when interpreting mosaicism. Sometimes when there is a mosaicism, one scientist says this is 40%, the other one says it's 60%.

And there are different points of view. And the artificial intelligence helps us to eliminate this individual interpretation. And it builds reference data that correlates the NGS results to the healthy live birth. So, when there is data and there is doubt, it links it to the data from previous implanted embryos and resulted in live birth. And we can consider it as robust and accurate interpretation. So, the conclusion is the artificial intelligence is a promising tool in embryo selection and in assistive reproduction technology as a general.

And it has been implemented in ART for a few years now. And it helps in decision-making, and it keeps improving by time. And we can have better prediction and better selection for embryos. And surely it saves time. Like in my lab, we have a rota for annotation of embryos in the time-lapse incubators.

So, when this system comes, it does auto-annotation and saves a lot of time. And it can do like several embryos together at the same time, and it can do it much faster than the human. And it's really an area of intensive research.

But it's still in early stages, needs a lot of development. There has been some development in recent few years, but we still need a lot of development to make better achievements in this field.

As AI continues to reshape embryo assessment and ART, it has become essential for aspiring embryologists and fertility professionals to receive structured, industry relevant training that covers these emerging achievements. In Medline Academics, the learner is able to gain a full understanding of embryology thanks to an education plan created by the experts in the field of IVF. The education will provide you with a strong theoretical basis and practical experience that will enable you to cope with the growing needs of IVF labs. While many prospective students compare the embryology course duration before enrolling, it is equally important to evaluate the quality of the curriculum, hands-on laboratory training, expert mentorship, and exposure to latest technologies used in assisted reproduction. Our Medline Academics program is concerned with offering our students an outcome-based educational experience that gives them the confidence, expertise, and scientific knowledge necessary for a successful career in clinical embryology and reproductive medicine.

Dr. Kamini Rao Hospitals, a renowned IVF hospital in Bangalore, is committed to integrating ART with personalized fertility care. With a team of experts, embryologists, and state-of-the art embryology laboratories, the hospital offers comprehensive evaluation and treatment for couples facing infertility. Through the integration of clinical expertise and innovations like advanced embryo evaluation, time-lapse photography, and precise laboratory protocols, Dr. Kamini Rao Hospitals is committed to providing patients with the best possible treatment results in addition to caring and patient-centered treatment. Clinical expertise, research, and innovations have made Dr. Kamini Rao Hospitals a preferred destination for people who are looking for advanced fertility solutions.