Niemeijer, Joshua und Ehrhardt, Jan und Uzunova, Hristina und Handels, Heinz (2024) TSynD: Targeted Synthetic Data Generation for Enhanced Medical Image Classification. 9th International Workshop, SASHIMI 2024, Held in Conjunction with MICCAI 2024, 2024-10-10, Marrakesh, Morocco. doi: 10.1007/978-3-031-73281-2_7. ISBN 978-3-031-73280-5.
PDF
7MB |
Kurzfassung
The usage of medical image data for the training of large-scale machine learning approaches is particularly challenging due to its scarce availability and the costly generation of data annotations, typically requiring the engagement of medical professionals. The rapid development of generative models enables us to tackle this problem by generating large amounts of realistic synthetic data for the training process. However, randomly choosing synthetic samples, might not be an optimal strategy. In this work, we investigate the targeted generation of synthetic training data, in order to improve the accuracy and robustness of image classification. Therefore, our approach aims to guide the generative model to synthesize data with high epistemic uncertainty, since large measures of epistemic uncertainty indicate underrepresented data points in the training set. During the image generation we feed images reconstructed by an auto encoder into the classifier and compute the mutual information over the class-probability distribution as a measure for uncertainty. We alter the feature space of the autoencoder through an optimization process with the objective of maximizing the classifier uncertainty on the decoded image. By training on such data we improve the performance and robustness against test time data augmentations and adversarial attacks on several classifications tasks.
elib-URL des Eintrags: | https://elib.dlr.de/207829/ | ||||||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Dokumentart: | Konferenzbeitrag (Vortrag) | ||||||||||||||||||||
Titel: | TSynD: Targeted Synthetic Data Generation for Enhanced Medical Image Classification | ||||||||||||||||||||
Autoren: |
| ||||||||||||||||||||
Datum: | 10 Oktober 2024 | ||||||||||||||||||||
Referierte Publikation: | Ja | ||||||||||||||||||||
Open Access: | Ja | ||||||||||||||||||||
Gold Open Access: | Nein | ||||||||||||||||||||
In SCOPUS: | Nein | ||||||||||||||||||||
In ISI Web of Science: | Nein | ||||||||||||||||||||
DOI: | 10.1007/978-3-031-73281-2_7 | ||||||||||||||||||||
ISBN: | 978-3-031-73280-5 | ||||||||||||||||||||
Status: | veröffentlicht | ||||||||||||||||||||
Stichwörter: | synthetic data generation, generalization, robustness, machine learning | ||||||||||||||||||||
Veranstaltungstitel: | 9th International Workshop, SASHIMI 2024, Held in Conjunction with MICCAI 2024 | ||||||||||||||||||||
Veranstaltungsort: | Marrakesh, Morocco | ||||||||||||||||||||
Veranstaltungsart: | internationale Konferenz | ||||||||||||||||||||
Veranstaltungsdatum: | 10 Oktober 2024 | ||||||||||||||||||||
HGF - Forschungsbereich: | Luftfahrt, Raumfahrt und Verkehr | ||||||||||||||||||||
HGF - Programm: | Verkehr | ||||||||||||||||||||
HGF - Programmthema: | Straßenverkehr | ||||||||||||||||||||
DLR - Schwerpunkt: | Verkehr | ||||||||||||||||||||
DLR - Forschungsgebiet: | V ST Straßenverkehr | ||||||||||||||||||||
DLR - Teilgebiet (Projekt, Vorhaben): | V - KoKoVI - Koordinierter kooperativer Verkehr mit verteilter, lernender Intelligenz, R - SynthBAD | ||||||||||||||||||||
Standort: | Braunschweig | ||||||||||||||||||||
Institute & Einrichtungen: | Institut für Verkehrssystemtechnik > Kooperative Systeme, BS | ||||||||||||||||||||
Hinterlegt von: | Niemeijer, Joshua | ||||||||||||||||||||
Hinterlegt am: | 28 Okt 2024 12:59 | ||||||||||||||||||||
Letzte Änderung: | 28 Okt 2024 12:59 |
Nur für Mitarbeiter des Archivs: Kontrollseite des Eintrags