Bi-Directional ConvLSTM U-Net with Densley Connected Convolutions (2019-08-31T00:00:00.000000Z)

TL;DR

This paper proposes an extension of U-Net, Bi-directional ConvLSTM U- net with Densely connected convolutions (BCDU-Net), for medical image segmentation, in which the full advantages of U -Net, bi- directional Conv lSTM (BConvL STM) and the mechanism of dense convolutions are taken.

Abstract

In recent years, deep learning-based networks have achieved state-of-the-art performance in medical image segmentation. Among the existing networks, U-Net has been successfully applied on medical image segmentation. In this paper, we propose an extension of U-Net, Bi-directional ConvLSTM U-Net with Densely connected convolutions (BCDU-Net), for medical image segmentation, in which we take full advantages of U-Net, bi-directional ConvLSTM (BConvLSTM) and the mechanism of dense convolutions. Instead of a simple concatenation in the skip connection of U-Net, we employ BConvLSTM to combine the feature maps extracted from the corresponding encoding path and the previous decoding up-convolutional layer in a non-linear way. To strengthen feature propagation and encourage feature reuse, we use densely connected convolutions in the last convolutional layer of the encoding path. Finally, we can accelerate the convergence speed of the proposed network by employing batch normalization (BN). The proposed model is evaluated on three datasets of: retinal blood vessel segmentation, skin lesion segmentation, and lung nodule segmentation, achieving state-of-the-art performance.

Authors

Sergio Escalera

31 papers

Reza Azad

4 papers

M. Fathy

2 papers

Bi-Directional ConvLSTM U-Net with Densley Connected Convolutions

TL;DR

Abstract

Authors

References27 items

Pyramid Dilated Deeper ConvLSTM for Video Salient Object Detection

Recurrent neural networks for aortic image sequence segmentation with sparse annotations

Attention U-Net: Learning Where to Look for the Pancreas

Recurrent Residual Convolutional Neural Network based on U-Net (R2U-Net) for Medical Image Segmentation

Deep Bidirectional and Unidirectional LSTM Recurrent Neural Network for Network-wide Traffic Speed Prediction

VoxResNet: Deep voxelwise residual networks for brain segmentation from 3D MR images

A Fixed-Point Model for Pancreas Segmentation in Abdominal CT Scans

Three-Dimensional CT Image Segmentation by Combining 2D Fully Convolutional Network with 3D Majority Voting

Densely Connected Convolutional Networks

The Importance of Skip Connections in Biomedical Image Segmentation

Brain MRI segmentation with patch-based CNN approach

3D U-Net: Learning Dense Volumetric Segmentation from Sparse Annotation

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

DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs

Skin lesion analysis toward melanoma detection: A challenge at the 2017 International symposium on biomedical imaging (ISBI), hosted by the international skin imaging collaboration (ISIC)

Deep MRI brain extraction: A 3D convolutional neural network for skull stripping

Segmenting Retinal Blood Vessels With Deep Neural Networks

ReSeg: A Recurrent Neural Network-Based Model for Semantic Segmentation

DeepOrgan: Multi-level Deep Convolutional Networks for Automated Pancreas Segmentation

Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcasting

U-Net: Convolutional Networks for Biomedical Image Segmentation

Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift

Trainable COSFIRE filters for vessel delineation with application to retinal images

Fully convolutional networks for semantic segmentation

Recurrent Convolutional Neural Networks for Scene Labeling

Ridge-based vessel segmentation in color images of the retina

A Cross-Modality Learning Approach for Vessel Segmentation in Retinal Images

Field of Study

Journal Information

Name

Page