基于改进YOLOv8s的肺结节检测算法OA

Lung cancer has become one of the common cancers with high mortality worldwide.Pulmonary nodules are the early manifestations of lung cancer.Aiming at the difficulty of pulmonary nodule detection in CT images,a pulmonary nodule detection algorithm based on improved YOLOv8s is proposed.The backbone network uses improved YOLOv8s,and uses Space-to-Depth Convolution to replace the traditional step-size convolution and pooling layer to avoid the loss of fine-grained information caused by step-size convolution and pooling layer when processing low-resolution images or small objects.The coordinate attention module is added to consider the inter-channel relationship and position information of the pulmonary nodule image,so that the model can locate and identify the target area more accurately.Focal Loss is used to replace the cross-entropy loss function to solve the problem of small number of sample labels in the dataset.The adaptive activation function can not only improve the stability of the network,but also improve the accuracy of the network.The LUNA16 dataset is used to verify the performance of the algorithm.The detection accuracy of the improved pulmonary nodule detection algorithm reaches 77.8%,which is 3.7%and 8.6%higher than that of Faster R-CNN and YOLOv8s detection algorithms,respectively.