Deep patch-based priors under a fully convolutional encoder-decoder architecture for interstitial lung disease segmentation

Abstract

Interstitial lung diseases (ILD) encompass a large spectrum of diseases sharing similarities in their physiopathology and computed tomography (CT) appearance. In this paper, we propose the adaption of a deep convolutional encoder-decoder (CED) that has shown high accuracy for image segmentation. Such architectures require annotation of the total region with pathological findings. This is difficult to acquire, due to uncertainty in the definition and extent of disease patterns and the need of significant human effort, especially for large datasets. Therefore, often current methods use patch-based implementations of convolutional neural networks, which however tend to produce spatially inhomogeneous segmentations due to their local contextual view. We exploit the advantages of both architectures by using the output of a patch-based classifier as a prior to a CED. Our method could advance the state-of-the-art in lung tissue segmentation using only a small number of newly annotated images.

AtlasNet: Multi-atlas Non-linear Deep Networks for Medical Image Segmentation

“…By extracting patches on our data (where different patterns are annotated as a single class) in the same way as in [1] we obtained 0.916 mean accuracy. In [12] a patch-based CNN was augmented with a deep encoder-decoder to exploit partial annotations. By applying AtlasNet on the same dataset as in [12], we increased the mean dice from 0.671 to 0.725.…”

“…In [12] a patch-based CNN was augmented with a deep encoder-decoder to exploit partial annotations. By applying AtlasNet on the same dataset as in [12], we increased the mean dice from 0.671 to 0.725.…”