OliverSchoppe

Chair for Computer Aided Medical Procedures & Augmented Reality
Lehrstuhl für Informatikanwendungen in der Medizin & Augmented Reality

Dr. Oliver Schoppe

Image-Based Biomedical Modeling Group (Prof. Dr. Bjoern Menze)
I finished my PhD in 2020 and I am not with the chair anymore

Oliver Schoppe

Deep Learning for clinical and pre-clinical research
Deep learning for clinical and pre-clinical research Image credits: NVIDIA;

Office
Center for Translational Cancer Research (TranslaTUM)
Klinikum rechts der Isar, Trogerstr 1, 81675 Munich

My TUM email address is not active anymore. In case of questions, please contact my colleagues at the chair or reach out to me via LinkedIn

Research Focus
My research interest is motivated by the tremendous potential impact of making Deep Learning-based computer vision applicable in everyday settings in clinical and pre-clinical research. This entails developing deep convolutional networks for detection, localization, and segmentation tasks in medical images of various modalities. Ensuring applicability in everyday settings typically requires elaborate strategies to deal with limited availability of annotated data, for instance making use of Transfer Learning or Data Augmentation.
I cooperate with Ali Ertürk's Lab at the Helmholtz Center Munich where I am leading the Artificial Intelligence Group. This work comprises research questions like automatic image segmentation and classification problems for general 3D anatomy, but also specific questions in the fields of stroke, dementia, and cancer.

DeepMACT - Deep learning based Metastasis Analysis in Cleared Tissue

The DeepMACT pipeline is an end-to-end procedure that enables highly automated detection and characterization of tumor micro-metastasis in large volumetric scans of whole cleared mice. This pipeline comprises two core technologies: DISCO tissue clearing and deep learning. For details, please refer to web resources and the article published in Cell (2019)




N-TV interview on DeepMACT Press coverage


AIMOS - AI-based Mouse Organ Segmentation

The AIMOS pipeline is an end-to-end procedure that enables fully automated segmentation of major organs (brain, lungs, heart, liver, kidneys, spleen, bladder, stomach, intestine) and the skeleton in large volumetric scans of whole mice. This method is independent of the imaging modality (we implemented it for light-sheet microscopy, native micro-CT, and contrast-enhanced micro-CT). In addition to the 3D segmentations for each organ, AIMOS can also provide a heatmap highlighting ambiguous image regions, allowing to localize and quantify inherent uncertainty. Besides the code and the pre-trained models, we also provide manually labeled datasets. For details, please refer to web resources and the article published in Nature Communications (2020)


TUM press statement on AIMOS Press coverage


Projects for students

We are inviting applications for several exciting research opportunities for Master students with our group. Please also feel free to reach out if you are interested in a Bachelor thesis, research project, etc. The projects are intended for students in Informatics, Electrical Engineering, or related fields. Please refer to the project descriptions below for more detailed information. Unfortunately, I cannot offer any new projects. If you are interested in student projects, please reach out to my IBBM colleagues directly.

We expect all students to have:
  • Prior theoretical knowledge on deep learning
  • Some practical experience with deep learning (PyTorch or TensorFlow)
  • Good Python coding skills
  • Ideally experience in computer vision / medical image analysis

Deep learning-based tumor detection in 3D scans

Deep transfer learning for 3D brain vessel segmentation

Deep learning-based tracing of single neurons and neural pathways

Distributed deep learning for large-scale medical image analysis

Deep learning-based 3D segmentation of prostate lesions

Analysis of CT scans from automotive OEM with deep learning

Deep learning-based multi-organ segmentation for MRI & CT

Deep learning-based multi-organ segmentation for 3D microscopy

Deep learning for multi-channel microscopy

Clinical application of deep learning for spine analysis

Machine learning analysis of sensor data for autonomous driving

NLP approaches to radiology report analysis

Image credits: Ali Ertürk; Anjany Sekuboyina; Fox; BMW AG; Shin et al.;

Background

  • 2018 - 2020: Ph.D. candidate at IBBM group (TU Munich) with Prof. Dr. Bjoern Menze and AI Group Leader at Helmholtz Center Munich
  • since 2015: Consultant with McKinsey & Co., Inc.
  • 2015: M.Sc. in Electrical Engineering and Information Technology from TU Munich
  • 2014-15: Research stays at Oxford University (neuroscience, artificial neural networks)
  • 2013: B.Sc. in Electrical Engineering and Information Technology from TU Munich
  • 2011-12: Visiting student at Harvard University & MIT (neuroscience, business, law)

Publications

  • S Zhao, M Todorov, R Cai, R al-Maskari, H Steinke, E Kemter, H Mai, Z Rong, M Warmer, K Stanic, O Schoppe, J Paetzold, B Gesierich, M Wong, T Huber, M Duering, O Bruns, B Menze, J Lipfert, V Puelles, E Wolf, I Bechmann, A Ertürk. Cellular and Molecular Probing of Intact Human Organs. Cell 2020

* joint first authorship
Google Scholar profile

Code



Edit | Attach | Refresh | Diffs | More | Revision r1.45 - 07 May 2021 - 06:42 - OliverSchoppe

Lehrstuhl für Computer Aided Medical Procedures & Augmented Reality    rss.gif