The development of neural networks enables computers to simulate human brain neurons to interpret images and identify image content accurately and quickly.Therefore,the technology of computer-aided medical diagnosis has certain research significance.Considering the detection of pneumonia CT as the classification of target detection,can extract the characteristics of pneumonia CT medical images through yolov5 to determine whether the patient has pneumonia,which can reduce the work of medical staff.Finally,the model detection effect is good,which can help doctors diagnose to a certain extent.