The computer-aided nodule detection system in CT image consists of the search sliding window, YOLOv2 architecture, 3-D CapsNet, skip connection,post-processing. First, the CT image is divided into numerous VOIs by sliding window. Second, a 3-D CapsNet based on YOLOv2 architectureskip connection is applied to the VOIs for classifying VOIs as nodulenot. Finally, the non-maximum suppression algorithm is performed to decide the final detection result.

The CADe system focuses on using 3-D CapsNet in YOLOv2 detection module to solve the problems of object rotationfeature shift. Simultaneously, the skip connection is also applied to network to overcome the vanishing-gradient problem for raising the accuracydowngrading the false positive rate. Compared to previous literatures, our system has better performanceuses fewer computation parameters.

It is necessary for the physician to spend more time to review image because of the enormous digitalized medical image. The CADe with deep learning can handle enormous digitalized medical image, reduce reviewing time,provide various nodule information for reducing the misdetection rate. As powerful learning capability, it can economize the cost in system designupdate system functionality by collecting image.