High-Order Graph Matching Based Feature Selection for Alzheimer's Disease Identification

Feng Liu, Heung-Il Suk, Chong-Yaw Wee, Huafu Chen, Dinggang Shen

Abstract

One of the main limitations of l 1-norm feature selection is that it focuses on estimating the target vector for each sample individually without considering relations with other samples. However, it’s believed that the geometrical relation among target vectors in the training set may provide useful information, and it would be natural to expect that the predicted vectors have similar geometric relations as the target vectors. To overcome these limitations, we formulate this as a graph-matching feature selection problem between a predicted graph and a target graph. In the predicted graph a node is represented by predicted vector that may describe regional gray matter volume or cortical thickness features, and in the target graph a node is represented by target vector that include class label and clinical scores. In particular, we devise new regularization terms in sparse representation to impose high-order graph matching between the target vectors and the predicted ones. Finally, the selected regional gray matter volume and cortical thickness features are fused in kernel space for classification. Using the ADNI dataset, we evaluate the effectiveness of the proposed method and obtain the accuracies of 92.17% and 81.57% in AD and MCI classification, respectively.
BibTeX
@inproceedings{liu2013high,
  author = {Feng Liu and Heung-Il Suk and Chong-Yaw Wee and Huafu Chen and Dinggang Shen},
  title = {{High-Order Graph Matching Based Feature Selection for Alzheimer's Disease Identification}},
  booktitle = {International Conference on Medical Image Computing and Computer-Assisted Intervention},
  year = {2013},
  doi = {10.1007/978-3-642-40763-5\_39},
}

Loading related content...

Figure