Automatic organ segmentation is a vital prerequisite of many clinical application in radiology. The anatomical variability of organs in the abdomen makes it difficult for many methods to obtain good segmentations for all organs. In this paper, we present a particular ensemble of convolutional neural networks, combining technologies that analyze the images with either a local or a global perspective. In particular, we implemented a cascade of models combining the advantages of using local and global processing. We have evaluated our proposed system on CT scan of 30 subjects in a nested cross-validation framework, showing a significant performance improvement if compared with state-of-the-art methods.

Multiple Organs Segmentation in Abdomen CT Scans Using a Cascade of CNNs

Sona, Diego
2019-01-01

Abstract

Automatic organ segmentation is a vital prerequisite of many clinical application in radiology. The anatomical variability of organs in the abdomen makes it difficult for many methods to obtain good segmentations for all organs. In this paper, we present a particular ensemble of convolutional neural networks, combining technologies that analyze the images with either a local or a global perspective. In particular, we implemented a cascade of models combining the advantages of using local and global processing. We have evaluated our proposed system on CT scan of 30 subjects in a nested cross-validation framework, showing a significant performance improvement if compared with state-of-the-art methods.
2019
978-3-030-30641-0
978-3-030-30642-7
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11582/319766
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
social impact