Best german team in Data Science Bowl 2017

The Data Science Bowl is the world’s highest endowed competition about 'Big Data'. This year’s challenge was to develop a software that recognizes lung cancer based on a single three-dimensional computed tomography (CT) scan – a task which is virtually impossible for radiologists. Alexander Wolf, Postdoc in the Machine Learning Group, Niklas Köhler and Julian Jungwirth, both from the Munich deep-learning startup Hellsicht and Moritz Berthold (master’s student at the ICB) developed an algorithm that can quickly and efficiently assess the CT images of the different layers achieving a much higher precision in the identification of small nodules than a radiologist. Another challenge for the team was to program the software to accurately distinguish between benign and cancerous lesions.

Alex Wolf and his colleagues finished as the best German team. Among the nearly 2000 participating teams, they achieved the seventh place.


For more information see the press release of the Helmholtz Zentrum München