Automatic Segmentation of Acute Ischemic Stroke From DWI Using 3-D Fully Convolutional DenseNets

Abstract

Acute ischemic stroke is recognized as a common cerebral vascular disease in aging people. Accurate diagnosis and timely treatment can effectively improve the blood supply of the ischemic area and reduce the risk of disability or even death. Understanding the location and size of infarcts plays a critical role in the diagnosis decision. However, manual localization and quantification of stroke lesions are laborious and time-consuming. In this paper, we propose a novel automatic method to segment acute ischemic stroke from diffusion weighted images (DWIs) using deep 3-D convolutional neural networks (CNNs). Our method can efficiently utilize 3-D contextual information and automatically learn very discriminative features in an end-to-end and data-driven way. To relieve the difficulty of training very deep 3-D CNN, we equip our network with dense connectivity to enable the unimpeded propagation of information and gradients throughout the network. We train our model with Dice objective function to combat the severe class imbalance problem in data. A DWI data set containing 242 subjects (90 for training, 62 for validation, and 90 for testing) with various types of acute ischemic stroke was constructed to evaluate our method. Our model achieved high performance on various metrics (Dice similarity coefficient: 79.13%, lesionwise precision: 92.67%, and lesionwise F1 score: 89.25%), outperforming the other state-of-the-art CNN methods by a large margin. We also evaluated the model on ISLES2015-SSIS data set and achieved very competitive performance, which further demonstrated its generalization capacity. The proposed method is fast and accurate, demonstrating a good potential in clinical routines.

Shape constrained fully convolutional DenseNet with adversarial training for multiorgan segmentation on head and neck CT and low‐field MR images

“…However, more layers can significantly impede the propagation of gradients, which is known as vanishing gradient . Deep residual network partly addresses this problem by introducing skip connections bypassing each residual blocks, but then the residual architecture can easily result in a large number of redundant features to impede the information flow as these connections are incorporated into networks by summation …”

“…31 Deep residual network 32 partly addresses this problem by introducing skip connections bypassing each residual blocks, but then the residual architecture can easily result in a large number of redundant features to impede the information flow as these connections are incorporated into networks by summation. 23 Dense blocks 33 are used in the analysis path of our segmentation network by hierarchically extracting the abstract representations of the input to facilitate gradients propagate to preceding layers and improve the network performance. Within each dense block, layers are directly connected with all their preceding layers by concatenation, which is shown in Fig.…”

“…Despite its theoretical appeal, deep learning segmentation in the 3D domain is still a largely unsolved problem. 3D CNN requires substantially more training samples than their 2D counterparts, thus easily suffer from overfitting, vanishing‐gradient problem, and high computational cost . To overcome these challenges, we developed a novel automated H&N OARs segmentation method that combines the fully convolutional residual network (FC‐ResNet) with a shape representation model (SRM).…”

“…3D CNN requires substantially more training samples than their 2D counterparts, thus easily suffer from overfitting, vanishing-gradient problem, and high computational cost. 23 To overcome these challenges, we developed a novel automated H&N OARs segmentation method that combines the fully convolutional residual network (FC-ResNet) with a shape representation model (SRM). The SRM network trained to capture the 3D OARs shape features were used to constrain the FC-ResNet.…”

Parallel Capsule Net for Ischemic Stroke Segmentation

“…The ischemic stroke which accounts for 87% of stroke cases is the most common cerebrovascular disease [1]. Ischemic stroke is caused by the obstruction of cerebral blood supply resulting in tissue hypoxia which progresses through several stages such as acute, subacute and chronic [2].…”