Detection of CT Pulmonary Nodule Based on Improved Yolo Using Attention Mechanism

Lung cancer is a leading malignant tumor in morbidity and mortality. Early diagnosis of lung cancer can effectively improve the prognosis. Pulmonary nodules are an essential feature of early symptoms of lung cancer. Therefore, the detection of pulmonary nodules is of great practical significance. On the basis of previous work, this paper proposes an improved Yolo network, which combines two attention mechanisms including CBAM (Convolutional Block Attention Module) and Multi-Head Self-Attention with yolov3 network to detect lung nodules in chest CT images. Experiments show that these attention mechanisms optimize the detection effect and make the performance of our method significantly better than the previous work.