V-Net：FullyConvolutionalNeuralNetworks

paper

需积分: 31 98 浏览量更新于2023-03-16 评论 1 收藏 4.15MB PDF 举报

身份认证购VIP最低享 7 折!

领优惠券(最高得80元）

资源详情

资源评论

资源推荐

V-Net: Fully Convolutional Neural Networks for

Volumetric Medical Image Segmentation

Fausto Milletari

, Nassir Navab

1,2

, Seyed-Ahmad Ahmadi

Computer Aided Medical Procedures, Technische Universit¨at M¨unchen, Germany

Computer Aided Medical Procedures, Johns Hopkins University, Baltimore, USA

Department of Neurology, Klinikum Grosshadern, Ludwig-Maximilians-Universit¨at

M¨unchen, Germany

Abstract. Convolutional Neural Networks (CNNs) have been recently

employed to solve problems from both the computer vision and medi-

cal image analysis ﬁelds. Despite their popularity, most approaches are

only able to process 2D images while most medical data used in clinical

practice consists of 3D volumes. In this work we propose an approach

to 3D image segmentation based on a volumetric, fully convolutional,

neural network. Our CNN is trained end-to-end on MRI volumes depict-

ing prostate, and learns to predict segmentation for the whole volume at

once. We introduce a novel objective function, that we optimise during

training, based on Dice coeﬃcient. In this way we can deal with situa-

tions where there is a strong imbalance between the number of foreground

and background voxels. To cope with the limited number of annotated

volumes available for training, we augment the data applying random

non-linear transformations and histogram matching. We show in our ex-

perimental evaluation that our approach achieves good performances on

challenging test data while requiring only a fraction of the processing

time needed by other previous methods.

1 Introduction and Related Work

Recent research in computer vision and pattern recognition has highlighted the

capabilities of Convolutional Neural Networks (CNNs) to solve challenging tasks

such as classiﬁcation, segmentation and object detection, achieving state-of-the-

art performances. This success has been attributed to the ability of CNNs to

learn a hierarchical representation of raw input data, without relying on hand-

crafted features. As the inputs are processed through the network layers, the

level of abstraction of the resulting features increases. Shallower layers grasp

local information while deeper layers use ﬁlters whose receptive ﬁelds are much

broader that therefore capture global information [19].

Segmentation is a highly relevant task in medical image analysis. Automatic

delineation of organs and structures of interest is often necessary to perform tasks

such as visual augmentation [10], computer assisted diagnosis [12], interventions

[20] and extraction of quantitative indices from images [1]. In particular, since

diagnostic and interventional imagery often consists of 3D images, being able to

arXiv:1606.04797v1 [cs.CV] 15 Jun 2016

Fig. 1. Slices from MRI volumes depicting prostate. This data is part of the

PROMISE2012 challenge dataset [7].

perform volumetric segmentations by taking into account the whole volume con-

tent at once, has a particular relevance. In this work, we aim to segment prostate

MRI volumes. This is a challenging task due to the wide range of appearance

the prostate can assume in diﬀerent scans due to deformations and variations of

the intensity distribution. Moreover, MRI volumes are often aﬀected by artefacts

and distortions due to ﬁeld inhomogeneity. Prostate segmentation is neverthe-

less an important task having clinical relevance both during diagnosis, where the

volume of the prostate needs to be assessed [13], and during treatment planning,

where the estimate of the anatomical boundary needs to be accurate [4,20].

CNNs have been recently used for medical image segmentation. Early ap-

proaches obtain anatomy delineation in images or volumes by performing patch-

wise image classiﬁcation. Such segmentations are obtained by only considering

local context and therefore are prone to failure, especially in challenging modal-

ities such as ultrasound, where a high number of mis-classiﬁed voxel are to be

expected. Post-processing approaches such as connected components analysis

normally yield no improvement and therefore, more recent works, propose to

use the network predictions in combination with Markov random ﬁelds [6], vot-

ing strategies [9] or more traditional approaches such as level-sets [2]. Patch-wise

approaches also suﬀer from eﬃciency issues. When densely extracted patches are

processed in a CNN, a high number of computations is redundant and therefore

the total algorithm runtime is high. In this case, more eﬃcient computational

schemes can be adopted.

Fully convolutional network trained end-to-end were so far applied only to 2D

images both in computer vision [11,8] and microscopy image analysis [14]. These

models, which served as an inspiration for our work, employed diﬀerent network

architectures and were trained to predict a segmentation mask, delineating the

structures of interest, for the whole image. In [11] a pre-trained VGG network

architecture [15] was used in conjunction with its mirrored, de-convolutional,

equivalent to segment RGB images by leveraging the descriptive power of the

features extracted by the innermost layer. In [8] three fully convolutional deep

neural networks, pre-trained on a classiﬁcation task, were reﬁned to produce

segmentations while in [14] a brand new CNN model, especially tailored to tackle

biomedical image analysis problems in 2D, was proposed.

In this work we present our approach to medical image segmentation that

leverages the power of a fully convolutional neural networks, trained end-to-end,

to process MRI volumes. Diﬀerently from other recent approaches we refrain

from processing the input volumes slice-wise and we propose to use volumetric

convolutions instead. We propose a novel objective function based on Dice coef-

ﬁcient maximisation, that we optimise during training. We demonstrate fast and

accurate results on prostate MRI test volumes and we provide direct comparison

with other methods which were evaluated on the same test data

2 Method

…

}

…

}

…

}

…

}

…

}

…

}

…

}

…

}

…

}

…

}

…

}

…

}

…

}

…

}

…

}

…

}

…

}

…

}

…

}

…

}

…

}

…

}

…

}

…

}

…

16 Channels

128 x 128 x 64

32 Channels

64 x 64 x 32

64 Channels

32 x 32 x 16

128 Channels

16 x 16 x 8

256 Channels

8 x 8 x 4

256 Channels

16 x 16 x 8

128 Channels

32 x 32 x 16

64 Channels

64 x 64 x 32

32 Channels

128 x 128 x 64

"Down" Conv.

"Up" Conv.

2 Ch. (Prediction)

128x128x64

1 Ch. (Input)

128x128x64

"Down" Conv.

}

Convolutional Layer

2x2 ﬁlters, stride: 2

"Up" Conv.

}

De-convolutional Layer

2x2 ﬁlters, stride: 2

Fine-grained features

forwarding

Convolution using a !

5x5x5 ﬁlter, stride: 1

PReLu non-linearity

Element-wise sum

Softmax

1x1x1

ﬁlter

Fig. 2. Schematic representation of our network architecture. Our custom imple-

mentation of Caﬀe [5] processes 3D data by performing volumetric convolutions.

Best viewed in electronic format.

In Figure 2 we provide a schematic representation of our convolutional neural

network. We perform convolutions aiming to both extract features from the data

Detailed results available on http://promise12.grand-challenge.org/results/

剩余10页未读，继续阅读

Kelisita

粉丝: 198
资源: 2

会员权益专享

V-Net：Fully Convolutional Neural Networks

评论0

会员权益专享

最新资源

V-Net：Fully Convolutional Neural Networks

评论0

VNet-Tensorflow:用于医学影像分割的V-Net架构的Tensorflow实现

fully convolutional reduced (atrous) VGGNet

快乐影院 v1.0

推荐20个以上比较好的目标检测模型github源码？

可以帮我找几个人工智能模型吗？

fully convolutional networks

著名的两阶段目标检测算法

pytorch对fashion mnist 分类

目前图像分割的流行算法

深度学习可以进行图像预处理、图像分割、特征识别吗

给我几篇关于人工智能与图像识别的文献

FCN（全卷积网络），CNN（卷积神经网络），RNN（循环神经网络），DeepLab系列，SegNet，U-Net简单介绍

基于深度学习的网络应用加密流量识别方法研究

实时语义分割有哪些算法

关键点检测可以用什么神经网络训练

基于深度学习的分割方法

maskRCNN分割

人工神经网络识别视频

使用模型进行图像语义分割,怎样修改训练网络

会员权益专享

最新资源