A manifold learning regularization approach to enhance 3D CT image-based lung nodule classification

ConclusionThe proposed MRC-DNN facilitates an accurate manifold learning approach for lung nodule classification based on 3D CT images. More importantly, MRC-DNN suggests a new and effective idea of enforcing regularization for network training, possessing the potential impact to a board range of applications.
Source: International Journal of Computer Assisted Radiology and Surgery - Category: Intensive Care Source Type: research