Endometriosis is a medical condition that affects approximately 6–10 percent of women worldwide, characterized by the growth of uterine-like tissue (endometrium) outside of the uterus, which commonly is painful and in severe cases can lead to infertility. The extent of endometriosis most frequently is determined by two established systems: the revised American Society for Reproductive Medicine (aSRM) score  and the European Enzian  classification. Both systems complement each other by enabling physicians to hand-calculate a severity score for different affected bodily regions and endometrium sizes. Based on retrieved results, a diagnosis is made and further proceedings are decided – in severe cases the abnormal tissue is surgically removed via laparoscopy.
We aim at alleviating the process of manual score calculation by developing a system providing automatic suggestions for endometriosis classification when processing laparoscopic media. For achieving this goal, we collaborate with leading endometriosis specialists, which provide us with sample endometriosis annotations including the anomaly’s shape, locality and severity. These annotations serve as a basis for computer-aided analyses and machine learning techniques in order to move closer towards automatic classifications and score calculations.