AI for Assisted Reproduction

Semen analysis is important in assessing the male partner in a subfertile couple. WHO has established guidelines for semen analysis, with sperm count, concentration, motility, vitality, and morphology as standard variables. Sperm concentration, morphology, and motility are shown to be associated with fertility. The methods constitute subjective assessments based on microscopic investigation, thus, standardization and quality assurance are of high importance. Nevertheless, there exists large between-laboratory variability for semen analysis, especially for sperm motility and morphology.

Computer assisted semen analysis (CASA) systems has improved since the first were introduced to the marked, but they are still not accurate and need improvement. Automatic classification of sperm morphology would be of significant assistance in semen analysis.


In this project, we aim to design and develop a system for analysing data collected in the process of assisted reproduction. The data consists of several different data types including videos, images and medical data collected in laboratories.
The main idea is to go beyond image based methods exploiting all possible data sources to provide predictions about sperm quality defined by morphology and motility. Algorithms to be explored are for example convolutional neural networks, recurrent neural networks and LSTMS but not limited to that.

Learning outcome

-Insight into advanced techniques of machine learning
-Working on a real world application
-Collaboration with researchers in the topic of machine learning, specifically deep learning
-Possibility to implement and research a novel approach


  • Programming
  • Motivation
  • Mathematics


Collaboration partners