Machine learning for analysis and diagnostic support in lung images Progress in machine vision based on „deep learning” combined
with large training datasets, e.g. ImageNet, have recently
attracted much attention. Some classification tasks on
natural medical images have already reached „super-human”
performance. With a giant hierarchical representation of
visual context, “deep learning” approaches can be applied on
medical images to improve the recognition and segmentation of
anatomical and pathological structures. However, suboptimal
techniques are being applied like brute force, i.e., the
complete scanning of an image. In this project, novel
approaches should be developed to support the recognition of
lung cancer and other diseases in medical lung images. | Project manager: Prof. Dr.-Ing. habil. Andreas Maier
Project participants: Sebastian Gündel, M. Eng.
Duration: 15.4.2017 - 15.4.2020
Sponsored by: Siemens Healhineers AG
Contact: Gündel, Sebastian Phone +49 9131 85 25246, Fax + 49 9131 85 27270, E-Mail: sebastian.guendel@fau.de
|