I am interested in interpretable machine learning algorithms for unsupervised anomaly detection.

Interests
  • Interpretable Machine Learning
  • Unsupervised Anomaly Detection
Education

  • Ph.D. Interpretable Machine Learning in Medical Image Analysis, ongoing

    Technical University of Munich

  • M.Sc. in Computer Science, 2018

    FAU Erlangen

  • B.Sc. in Computer Science, 2015

    FAU Erlangen

Research

PHANES
Reversing the Abnormal: Pseudo-Healthy Generative Networks for Anomaly Detection.
PHANES
RA
Moving beyond hyperintensity thresholding. This work analyzes the challenges and outlines opportunities for advancing the field of unsupervised anomaly detection.
RA
MorphAEus
Novel deformable auto-encoders for unsupervised outlier detection.
MorphAEus
See all

Publications

Cosmin I. Bercea, Benedikt Wiestler, Daniel Rückert, Julia A Schnabel (2023). Generalizing Unsupervised Anomaly Detection: Towards Unbiased Pathology Screening. In MIDL.

PDF Cite Project

Cosmin I. Bercea, Benedikt Wiestler, Daniel Rueckert, Julia Schnabel (2023). Reversing the Abnormal: Pseudo-Healthy Generative Networks for Anomaly Detection. In arxiv.

PDF Cite Project

Cosmin I. Bercea, Benedikt Wiestler, Daniel Rückert, Shadi Albarqouni (2022). Federated Disentangled Representation Learning for Unsupervised Brain Anomaly Detection. In Nature Machine Intelligence.

PDF Cite Code Project Video

Cosmin I. Bercea, Daniel Rueckert, Julia Schnabel (2022). What do we learn? Debunking the Myth of Unsupervised Outlier Detection. In arxiv.

PDF Cite Project

Cosmin I. Bercea, Benedikt Wiestler, Daniel Rückert, Shadi Albarqouni (2020). FedDis: Disentangled Federated Learning for Unsupervised Brain Pathology Segmentation. In arxiv.

PDF Cite Code Project Video

See all publications

Teaching

If you are interested in BSc/MSc projects in unsupervised anomaly detection please contact me and attach a motivation letter, transcript of academic records and CV.

*
All Bachelor Thesis Master Thesis Seminars
Deep Medical Anomaly Segmentation

Deep Medical Anomaly Segmentation

Summer semester 2023. TUM Informatics. Master Seminar.

Unsupervised Anomaly Detection for Medical Imaging (f/m/x)

Unsupervised Anomaly Detection for Medical Imaging (f/m/x)

Bachelor/Master Thesis)