Share Email Print

Proceedings Paper

Orientation regression in hand radiographs: a transfer learning approach
Author(s): Ivo M. Baltruschat; Axel Saalbach; Mattias P. Heinrich; Hannes Nickisch; Sascha Jockel
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

Most radiologists prefer an upright orientation of the anatomy in a digital X-ray image for consistency and quality reasons. In almost half of the clinical cases, the anatomy is not upright orientated, which is why the images must be digitally rotated by radiographers. Earlier work has shown that automated orientation detection results in small error rates, but requires specially designed algorithms for individual anatomies. In this work, we propose a novel approach to overcome time-consuming feature engineering by means of Residual Neural Networks (ResNet), which extract generic low-level and high-level features, and provide promising solutions for medical imaging. Our method uses the learned representations to estimate the orientation via linear regression, and can be further improved by fine-tuning selected ResNet layers. The method was evaluated on 926 hand X-ray images and achieves a state-of-the-art mean absolute error of 2.79°.

Paper Details

Date Published: 2 March 2018
PDF: 8 pages
Proc. SPIE 10574, Medical Imaging 2018: Image Processing, 105741W (2 March 2018); doi: 10.1117/12.2291620
Show Author Affiliations
Ivo M. Baltruschat, Univ. zu Lübeck (Germany)
Philips Medical Systems DMC GmbH (Germany)
Axel Saalbach, Philips GmbH Innovative Technologies (Germany)
Mattias P. Heinrich, Univ. zu Lübeck (Germany)
Hannes Nickisch, Philips GmbH Innovative Technologies (Germany)
Sascha Jockel, Philips Medical Systems DMC GmbH (Germany)

Published in SPIE Proceedings Vol. 10574:
Medical Imaging 2018: Image Processing
Elsa D. Angelini; Bennett A. Landman, Editor(s)

© SPIE. Terms of Use
Back to Top