Intelligent Detection Model of Lung Nodules in Medical CT Images Based on Deep Learning

Aiming at the problem that the traditional diagnosis of lung nodules relies on doctors' experience, which is time-consuming and prone to misdiagnosis, this study proposes an intelligent detection method for lung nodules based on the YOLOv11 deep learning framework. Through the training and validation of 2356 clinical CT images (1645 in the training set, 474 in the validation set, and 237 in the test set), the model achieves breakthrough performance in the lung nodule detection task: the mean intersection ratio (MIoU) reaches 84.23%, the accuracy (A) 98.56%, and the precision (P) 78.1%. Experiments show that YOLOv11 significantly optimizes the lung nodule localization accuracy and false detection suppression ability through the dynamic confidence threshold mechanism and C2PSA feature fusion module. The results provide highly reliable technical support for the early diagnosis of lung cancer, which is superior to the existing YOLO series models (11% MIoU improvement over YOLOv5).