There are many active research projects accessing and applying shared ADNI data. Use the search above to find specific research focuses on the active ADNI investigations. This information is requested annually as a requirement for data access.
| Principal Investigator | |
| Principal Investigator's Name: | TIAN XIONG |
| Institution: | University of Shanghai for Science and Technology |
| Department: | Electronic Engeneering |
| Country: | |
| Proposed Analysis: | Abstract—It has been a hot topic to detect Alzheimer's disease (AD) from MRI and other neuroimaging data by machine learning in recent years. A large number of deep learning models have been developed, such as alexnet, VGg net, Gan and so on. The common limitation of these models is that they rely on a large number of labeled training images and need deeper network structure to achieve higher accuracy. It is challenging to obtain a large number of labeled data in the field of medical image. Training a suitable medical image classification model from scratch also requires a lot of resources. To solve these problems, in this paper, we try to use convolutional network (CNN) transfer learning for image classification training, select CNN network model, fine tune the network weight layer by layer, modify the last layer of CNN for training on MRI image. In order to reduce the size of training data, we use image entropy to select the most informative slice. Through the experiments on ADNI data sets, compared with other contemporary methods, we achieve 99.89% on the classification of Alzheimer's disease (AD), Normal Cognitive (NC) and Mild Cognitive Impairment (MCI). Attractively, class activation mapping (CAM) is employed to visualize brain regions related to AD, which further gives the targeted information for AD diagnosis. |
| Additional Investigators |
