Voxel Deconvolutional Networks for 3D Brain Image Labeling
Voxel Deconvolutional Networks for 3D Brain Image Labeling
复制标题
DOI：
10.1145/3219819.3219974
复制
发表时间：
2018-07
期刊：
影响因子：
--
通讯作者：
Yongjun Chen;Hongyang Gao;Lei Cai;Min Shi-;D. Shen;Shuiwang Ji
中科院分区：
文献类型：
--
作者：
Yongjun Chen;Hongyang Gao;Lei Cai;Min Shi-;D. Shen;Shuiwang Ji
深度学习方法在像素预测任务中取得了巨大成功。最流行的方法之一是采用编码器-解码器网络，其中反卷积层用于对特征图进行上采样。然而，去卷积层的一个关键限制是它受到棋盘伪影问题的影响，这会损害预测精度。这是由输出特征图上相邻像素之间的独立性引起的。先前的工作仅解决了2D空间中去卷积层的棋盘伪影问题。由于生成解卷积层所需的中间特征图的数量随着维度呈指数增长，因此在更高维度中解决这个问题更具挑战性。在这项工作中，我们提出了体素去卷积层（VoxelDCL），以解决棋盘状伪影问题的去卷积层在3D空间。我们还提供了一个有效的方法来实现体素DCL。为了证明体素DCL的有效性，我们建立了四种变化的体素去卷积网络（体素DCN）的基础上的U-Net架构与体素DCL。我们使用ADNI和LONI LPBA 40数据集将我们的网络应用于体积脑图像标记任务。实验结果表明，所提出的iVoxelDCNa算法在所有实验中都取得了较好的性能. ADNI数据集上的骰子比例达到83.34%，LONI LPBA 40数据集上的骰子比例达到79.12%，分别较基线提高1.39%和2.21%。此外，我们提出的所有VoxelDCN变体在上述数据集上的性能都优于基线方法，这证明了我们方法的有效性。
Deep learning methods have shown great success in pixel-wise prediction tasks. One of the most popular methods employs an encoder-decoder network in which deconvolutional layers are used for up-sampling feature maps. However, a key limitation of the deconvolutional layer is that it suffers from the checkerboard artifact problem, which harms the prediction accuracy. This is caused by the independency among adjacent pixels on the output feature maps. Previous work only solved the checkerboard artifact issue of deconvolutional layers in the 2D space. Since the number of intermediate feature maps needed to generate a deconvolutional layer grows exponentially with dimensionality, it is more challenging to solve this issue in higher dimensions. In this work, we propose the voxel deconvolutional layer (VoxelDCL) to solve the checkerboard artifact problem of deconvolutional layers in 3D space. We also provide an efficient approach to implement VoxelDCL. To demonstrate the effectiveness of VoxelDCL, we build four variations of voxel deconvolutional networks (VoxelDCN) based on the U-Net architecture with VoxelDCL. We apply our networks to address volumetric brain images labeling tasks using the ADNI and LONI LPBA40 datasets. The experimental results show that the proposed iVoxelDCNa achieves improved performance in all experiments. It reaches 83.34% in terms of dice ratio on the ADNI dataset and 79.12% on the LONI LPBA40 dataset, which increases 1.39% and 2.21% respectively compared with the baseline. In addition, all the variations of VoxelDCN we proposed outperform the baseline methods on the above datasets, which demonstrates the effectiveness of our methods.