A bit more on semantic segmentation, now 3D

{V-Net: Fully Convolutional Neural Networks for Volumetric Medical Image Segmentation
}

--> Link / authors http://arxiv.org/abs/1606.04797, Fausto Milletari / Nassir Navab / Seyed-Ahmad Ahmadi
--> Essence:
(0) Essentially applies UNet to 3D with a custom DICE based loss
(1) Architecture - https://goo.gl/Yn2BGb - basically UNet with 3D convolutions. Upsampling / downsampling - https://goo.gl/VtXrXy
(2) PReLu (no ablation test)
(3) Receptive fields of layers - https://goo.gl/FGwDCF
(4) 3D DICE loss - https://goo.gl/SqrK93 (wo BCE?)
--> The paper does not use all the juice possible - hacky transfer learning (obvious idea - just stacking Imagenet filters), CLR, LinkNet architectures, etc
--> Looks like a good baseline / reference

{An application of cascaded 3D fully convolutional networks for medical image segmentation
}

--> http://arxiv.org/abs/1803.05431, a group of Japanese researchers
--> Essence:
(0) 2 stage 3D UNet, ablation test against 2D FCNs
(1) Loss - 3D cross-entropy
(2) Transfer learning - it works for other datasets, give a mild boost (1-3 %)
(3) 80-90% DICE, varies by organ
(4) weights downloadable https://github.com/holgerroth/3Dunet_abdomen_cascade (Caffe...)
--> Essentially a 2 stage process is dictated by memory considerations:
(0) Pipeline https://goo.gl/wZwF3X

In the long run transfer learning may rule, but here legal limitations may slow down this process.

#deep_learning
#medical_imaging