Unsupervised Detection of Disturbances in 2D Radiographs

  • We present a method based on a generative model for detection of disturbances like prosthesis, screws, zippers, and metals in 2D radiographs. The generative model is trained in an unsupervised fashion using clinical radiographs as well as simulated data both containing no disturbances. Our approach employs a latent space consistency loss which has the benefit that identity similarity is enforced to reconstruct X-rays without disturbances. In order to detect images with disturbances, an anomaly score is computed employing also the Fréchet distance between an input X-ray and the one reconstructed using our generative model. Validation was performed using clinical pelvis radiographs. We achieved an AUC of 0.77 and 0.83 with clinical and synthetic data, respectively. The results demonstrated a good accuracy of our method for detecting outliers as well as the advantage of utilizing synthetic data.

Export metadata

Additional Services

Share in Twitter Search Google Scholar
Metadaten
Author:Laura Estacio, Moritz Ehlke, Alexander TackORCiD, Eveling Castro-Gutierrez, Hans Lamecker, Rensso Mora, Stefan ZachowORCiD
Document Type:Article
Parent Title (English):IEEE 18th International Symposium on Biomedical Imaging (ISBI 2021)
First Page:367
Last Page:370
Year of first publication:2021