Pulmonary Nodule Detection Using 3-D Residual U-Net Oriented Context-Guided Attention and Multi-Branch Classification Network

Abstract

Accurate detection of pulmonary nodules on chest computed tomography scans is crucial to early diagnosis of lung cancer. To address the thorn problems on low detection sensitivity and high falsepositive rate caused by heterogeneity and morphological complexity of 3-D nodule features, a computeraided detection system is developed to increase the detection sensitivity and classification accuracy of pulmonary nodules. The contributions include: (1) Nodule candidate detection: 3-D Residual U-Net model is improved to detect candidate nodules, which constructs 3-D context-guided module to extract local and global nodule features by setting the dilated convolution with different dilation rates. Furthermore, channel attention mechanism is used to dynamically adjust the channel features, which enhances the generalization and expression ability of the detection-network to effectively learn 3-D spatial context features. (2) False-positive reduction: multi-branch classification network is designed for multi-task learning. Image reconstruction task is performed to retain more microscopic nodules information from convolutional neural network (CNN) hierarchy. Moreover, each branch deals with the feature map at corresponding depth layers, and then all branches' feature maps are combined together to perform nodule classification task. Numerous experimental results show that the proposed system is perfectly qualified for pulmonary nodules detection on Lung Nodules Analysis 2016 dataset, which achieves detection sensitivity up to 94.0% and competition performance metric (CPM) score up to 0.959.

The diagnosis performance of convolutional neural network in the detection of pulmonary nodules: a systematic review and meta-analysis

“…We read the remaining 253 full texts, and 227 studies were excluded for the following reasons: (a) 41 were benign and malignant diagnoses; (b) 156 only included sensitivity results; (c) 22 were without the relevant data; and (d) eight were review articles. Finally, 26 studies (19–44), including 2,391,702 ROIs, were included in the quantitative assessment and final combinatorial analysis. The flow diagram is shown in Fig.…”

MENet: A Mitscherlich function based ensemble of CNN models to classify lung cancer using CT scans

“…Zhao et al [ 32 ] proposed a forward and backward GAN with multi-scale VGG16. Yuan et al [ 33 ] proposed a CAD method to enhance the detection of pulmonary nodules on CT scans. Their proposed method utilized a 3D Residual U-Net model along with a multi-branch classification network to achieve a high detection sensitivity of 94.0% and a competition performance metric (CPM) score of 0.959 on the LUNA 2016 dataset through multi-task learning.…”

3D Efficient Multi-Task Neural Network for Knee Osteoarthritis Diagnosis Using MRI Scans: Data From the Osteoarthritis Initiative

“…Unlike 2D CNN, this allows a more comprehensive analysis of the anatomical structures or abnormalities which potentially improves the automated tasks' performance. The robustness of this transition provides new research directions for future research [5]. Despite the advancement of 3D CNN in computer-aided diagnosis, there is still a gap of knowledge in 3D CNN applications.…”