SHAMANN: Shared Memory Augmented Neural Networks

Bercea CI, Pauly O, Maier A, Ghesu FC (2019)


Publication Type: Conference contribution

Publication year: 2019

Journal

Publisher: Springer Verlag

Book Volume: 11492 LNCS

Pages Range: 830-841

Conference Proceedings Title: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

Event location: Hong Kong CN

ISBN: 9783030203504

DOI: 10.1007/978-3-030-20351-1_65

Abstract

Current state-of-the-art methods for semantic segmentation use deep neural networks to learn the segmentation mask from the input image signal as an image-to-image mapping. While these methods effectively exploit global image context, the learning and computational complexities are high. We propose shared memory augmented neural network actors as a dynamically scalable alternative. Based on a decomposition of the image into a sequence of local patches, we train such actors to sequentially segment each patch. To further increase the robustness and better capture shape priors, an external memory module is shared between different actors, providing an implicit mechanism for image information exchange. Finally, the patch-wise predictions are aggregated to a complete segmentation mask. We demonstrate the benefits of the new paradigm on a challenging lung segmentation problem based on X-Ray images, as well as on two synthetic tasks based on MNIST. On the X-Ray data, our method achieves state-of-the-art accuracy with a significantly reduced model size, 3–5 times compared to reference methods. In addition, we reduce the number of failure cases by at least half.

Authors with CRIS profile

Involved external institutions

How to cite

APA:

Bercea, C.I., Pauly, O., Maier, A., & Ghesu, F.C. (2019). SHAMANN: Shared Memory Augmented Neural Networks. In Siqi Bao, Albert C.S. Chung, James C. Gee, Paul A. Yushkevich (Eds.), Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (pp. 830-841). Hong Kong, CN: Springer Verlag.

MLA:

Bercea, Cosmin I., et al. "SHAMANN: Shared Memory Augmented Neural Networks." Proceedings of the 26th International Conference on Information Processing in Medical Imaging, IPMI 2019, Hong Kong Ed. Siqi Bao, Albert C.S. Chung, James C. Gee, Paul A. Yushkevich, Springer Verlag, 2019. 830-841.

BibTeX: Download