The in Vitro Fertilization

Since 1978 the In Vitro Fertilization is defined as the inability of the couple to conceive for the at least one year with a timely sexual intercourse without any birth control [2]. The IVF process involves collection of embryo which has to be inseminated by the sperm under clinical conditions. These fertilized embryos are under observation at least for the period of 2-5 days. The embryo which is good for implantation will be selected by the embryologists and then it will be transferred to the woman’s womb either at day 2 or at day 5. To check for the viable embryo is a tedious process which involves experts such as embryologists to be present physically. But still the success remains 20-25%, because of lack of identifying a potential embryo. To have the possibility of pregnancy multiple embryo will be transferred to women’s womb. This multiple transfer will be complicated for both mother and baby. Several investigators have been looking for various solutions to clearly identify and transfer single embryo.

Overall treatment of the IVF depends on individual cycle response, patient’s ability to accept, clinical aspects, embryo viability, equipment technology. Personal experiences of the individuals as patients, clinicians and embryologists.

The machine learning techniques can be applied to the IVF process to increase the efficiency of the selection. A model can be designed to evaluate these embryos for the implantation process, which will train itself with given parameters by providing automated decision support to embryologists when need exists. On the contrary to the appearance and significance of decision support systems in IVF process, the related literature is limited. Arti?cial Neural Networks (ANN), Convolutional Neural Networks (CNN), ReLU network classi?ers and also the prediction models are used to the neural network to obtain the accurate outcome in IVF treatment. Machine Learning techniques are the prediction models in which the network learns to accomplish classification of digital images or any tasks directly from given set of images, text, or sound.

The medical data obtained will be in text form. Retrieval of such data becomes complex. Machine learning is usually implemented using neural network architecture, here in this paper the machine learning model is done by training the network through dataset obtained by several hospitals. Once the data is trained the analysis on any image can be obtained very easily.

Generally the machine learning networks contain several connected layers of convolutional neural networks which can be operated on to classifiers. Machine learning techniques are giving better result compared to Hugh’s transform algorithm and Multi scale Vesselness filtering. Applying these techniques has improved the performance by 96.7% and training the network is faster than the previous algorithms. Recognizing viability of human embryos from microscopic images is an extremely tedious process that is susceptible to error and subject to intra- and inter-individual unpredictability.

Automating classification of these embryo images will have the bene?t of reducing time and cost, minimizing errors, and improving outcome, consistency of results between individuals and clinics. Several techniques have been discussed in literature to ease the process of automation taking into consideration of day2 as well as day3 embryo images. However, grading the embryos based on the cell division, size of the cell and the fragments present becomes difficult because of constraints in the imaging process. Like, the exposure time (embryos are sensitive to the temperature), the light intensity variation and the transparency of the specimen all cause variations in the image. Embryo quality assessment based on Blastomere circle and grading do not yield sufficiently reliable classification.