mmsegmentation/configs/unet

U-Net: Convolutional Networks for Biomedical Image Segmentation

Introduction

Official Repo

Code Snippet

Abstract

There is large consent that successful training of deep networks requires many thousand annotated training samples. In this paper, we present a network and training strategy that relies on the strong use of data augmentation to use the available annotated samples more efficiently. The architecture consists of a contracting path to capture context and a symmetric expanding path that enables precise localization. We show that such a network can be trained end-to-end from very few images and outperforms the prior best method (a sliding-window convolutional network) on the ISBI challenge for segmentation of neuronal structures in electron microscopic stacks. Using the same network trained on transmitted light microscopy images (phase contrast and DIC) we won the ISBI cell tracking challenge 2015 in these categories by a large margin. Moreover, the network is fast. Segmentation of a 512x512 image takes less than a second on a recent GPU. The full implementation (based on Caffe) and the trained networks are available at this http URL.

UNet (MICCAI'2016/Nat. Methods'2019)

@inproceedings{ronneberger2015u,
  title={U-net: Convolutional networks for biomedical image segmentation},
  author={Ronneberger, Olaf and Fischer, Philipp and Brox, Thomas},
  booktitle={International Conference on Medical image computing and computer-assisted intervention},
  pages={234--241},
  year={2015},
  organization={Springer}
}

Results and models

DRIVE

Method	Backbone	Image Size	Crop Size	Stride	Lr schd	Mem (GB)	Inf time (fps)	Dice	config	download
FCN	UNet-S5-D16	584x565	64x64	42x42	40000	0.680	-	78.67	config	model | log
PSPNet	UNet-S5-D16	584x565	64x64	42x42	40000	0.599	-	78.62	config	model | log
DeepLabV3	UNet-S5-D16	584x565	64x64	42x42	40000	0.596	-	78.69	config	model | log

STARE

Method	Backbone	Image Size	Crop Size	Stride	Lr schd	Mem (GB)	Inf time (fps)	Dice	config	download
FCN	UNet-S5-D16	605x700	128x128	85x85	40000	0.968	-	81.02	config	model | log
PSPNet	UNet-S5-D16	605x700	128x128	85x85	40000	0.982	-	81.22	config	model | log
DeepLabV3	UNet-S5-D16	605x700	128x128	85x85	40000	0.999	-	80.93	config	model | log

CHASE_DB1

Method	Backbone	Image Size	Crop Size	Stride	Lr schd	Mem (GB)	Inf time (fps)	Dice	config	download
FCN	UNet-S5-D16	960x999	128x128	85x85	40000	0.968	-	80.24	config	model | log
PSPNet	UNet-S5-D16	960x999	128x128	85x85	40000	0.982	-	80.36	config	model | log
DeepLabV3	UNet-S5-D16	960x999	128x128	85x85	40000	0.999	-	80.47	config	model | log

HRF

Method	Backbone	Image Size	Crop Size	Stride	Lr schd	Mem (GB)	Inf time (fps)	Dice	config	download
FCN	UNet-S5-D16	2336x3504	256x256	170x170	40000	2.525	-	79.45	config	model | log
PSPNet	UNet-S5-D16	2336x3504	256x256	170x170	40000	2.588	-	80.07	config	model | log
DeepLabV3	UNet-S5-D16	2336x3504	256x256	170x170	40000	2.604	-	80.21	config	model | log