Automatic Classification of Proximal Femur Fractures Based on Attention Models

Anees Kazi, Shadi Albarqouni, Amelia Jiménez Sánchez, Sonja Kirchhoff, Peter Biberthaler, Nassir Navab, Diana Mateus

Publikation: Konference artikel i Proceeding eller bog/rapport kapitelKonferencebidrag i proceedingsForskningpeer review


We target the automatic classification of fractures from clinical X-Ray images following the Arbeitsgemeinschaft Osteosynthese (AO) classification standard. We decompose the problem into the localization of the region-of-interest (ROI) and the classification of the localized region. Our solution relies on current advances in multi-task end-to-end deep learning. More specifically, we adapt an attention model known as Spatial Transformer (ST) to learn an image-dependent localization of the ROI trained only from image classification labels. As a case study, we focus here on the classification of proximal femur fractures. We provide a detailed quantitative and qualitative validation on a dataset of 1000 images and report high accuracy with regard to inter-expert correlation values reported in the literature.
TitelMLMI 2017: Machine Learning in Medical Imaging
Publikationsdato7 sep. 2017
ISBN (Trykt)978-3-319-67388-2
ISBN (Elektronisk)978-3-319-67389-9
StatusUdgivet - 7 sep. 2017
Udgivet eksterntJa


Dyk ned i forskningsemnerne om 'Automatic Classification of Proximal Femur Fractures Based on Attention Models'. Sammen danner de et unikt fingeraftryk.