Research on Pulmonary Nodule Detection Method Based on YOLO Algorithm

This paper presents a method for pulmonary nodule detection in CT images based on an improved YOLOv8 framework. Against the backdrop of current deep learning technologies, automated detection and segmentation of medical images are crucial for enhancing diagnostic speed and accuracy. We combine the YOLOv8 object detection framework with Receptive Fields Block (RFB) modules to enhance feature extraction capabilities. Additionally, we introduce Inner-IoU as a loss function to optimize the detector's localization accuracy. Experimental results demonstrate significant performance improvement in pulmonary nodule detection tasks on CT images, thereby providing substantial support for the automation and intelligence of medical image processing.