Abstract:Automatic detection of pulmonary nodules is critical for the early diagnosis and prevention of lung cancer. Computed tomography (CT) is an effective and economical lung cancer detection method. In CT images, the size and shape of pulmonary nodules appear different, and some nodules appear similar to the surrounding tissues. Therefore, the automatic localization of pulmonary nodules in CT images is a challenging task. An attention-embedded three-dimensional convolutional neural network is proposed for pulmonary nodule detection in the current study. Specifically, 1) channel-spatial attention guides 3D ResNet to down sample the input 3D CT patch. The channel pays attention to important features and the space to the region of interest. The two form a complementary feature extraction mechanism to effectively help the global flow of information in the network and refine the feature mapping to extract the nodule context features. 2) The channel-spatial attention module changes the fusion model of the feature pyramid, adaptively adjusts the pixel-level weight between features and extracts multi-scale representative node features. 3) The deep separable convolution is used to replace the standard convolution of ResNet, reducing the time cost and improving the efficiency of model training on the premise of ensuring the model’s performance. 4) To adapt the distribution of nodule scale, different characteristic layers correspond to two sizes of anchors. Under the condition of ensuring the detection rate of nodules, the number of anchor frames is reduced, and the network sensitivity is improved. Finally, several ablation experiments are carried out using the LUNA16 dataset. The results revealed that the attention-guided network could extract the multi-scale representative features of nodules, and the average sensitivity was 97.7%. Additionally, the CMP score reached 0.912. The extensive experiments demonstrate that the proposed approach can effectively improve the detection sensitivity and control the number of false positive nodules, which has clinical application value and a certain reference value.

The Detection and Segmentation of Pulmonary Nodules Based on U-Net.

Modified U-Net block network for lung nodule detection

Identification of Pulmonary Nodules Via CT Images with Hierarchical Fully Convolutional Networks.

Lung Nodule Segmentation and Recognition Algorithm Based on Multiposition U-Net

A Novel Pulmonary Nodule Detection Model Based on Multi-Step Cascaded Networks

Automated Pulmonary Nodule Detection in CT Images Using Deep Convolutional Neural Networks.

Pulmonary Nodule Detection in CT Images Via Deep Neural Network: Nodule Candidate Detection

Design of Lung Nodules Segmentation and Recognition Algorithm Based on Deep Learning

Pulmonary Nodule Detection Based on CT Images Using Convolution Neural Network

A Deep Learning Model of Automatic Detection of Pulmonary Nodules Based on Convolution Neural Networks (CNNs).

Pulmonary nodules detection based on multi-scale attention networks

Fast and Fully-Automated Detection and Segmentation of Pulmonary Nodules in Thoracic CT Scans Using Deep Convolutional Neural Networks

A Deep-Learning Based Automatic Pulmonary Nodule Detection System

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

Automatic Lung Nodule Detection Using a 3D Deep Convolutional Neural Network Combined with a Multi-Scale Prediction Strategy in Chest CTs

Attention-Guided Feature Extraction and Multiscale Feature Fusion 3D ResNet for Automated Pulmonary Nodule Detection.

ResDSda_U-Net: A Novel U-Net-Based Residual Network for Segmentation of Pulmonary Nodules in Lung CT Images

Lung Nodule Classification of CT Images Based on the Deep Learning Algorithms

Automated Detection and Classification for Early Stage Lung Cancer on CT Images Using Deep Learning

A Image Preprocessing Method Based on Visual Attention Mechanism for Deep Learning Detection of Pulmonary Nodules

Pulmonary CT Nodules Segmentation Using an Enhanced Square U-Net with Depthwise Separable Convolution