Bone suppression in chest x-rays is an important processing step that can often improve visual detection of lung pathologies hidden under ribs or clavicle shadows. Current diagnostic imaging protocol does not include hardware-based bone suppression, hence the need for a software-based solution. This paper evaluates various deep learning models adapted for bone suppression task, namely, we implemented several state-of-the-art deep learning architectures: convolution autoencoder, U-net, FPN, cGAN; augmented them with domain-specific denoising techniques, such as wavelet decomposition, with the aim to identify the optimal solution for chest x-ray analysis. Our results show that wavelet decomposition does not improve the rib suppression, “skip connections” modification outperforms baseline autoencoder approach with and without the usage of the wavelet decomposition, the residual models are trained faster than plain models and achieve higher validation scores.
CITATION STYLE
Sirazitdinov, I., Kubrak, K., Kiselev, S., Tolkachev, A., Kholiavchenko, M., & Ibragimov, B. (2020). Evaluation of Deep Learning Methods for Bone Suppression from Dual Energy Chest Radiography. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 12396 LNCS, pp. 247–257). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-61609-0_20
Mendeley helps you to discover research relevant for your work.