.png){kind=icon}
.png){kind=icon}
Return
Nested cross-validation based adaptive sparse representation algorithm and its application to pathological brain classification
.png){kind=icon}
PRE
AI
DOI：10.1016/j.eswa.2018.07.039.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Brain disease such as brain tumor, Alzheimer's disease, etc. is a major public health problem, and the main cause of death worldwide. Expert systems are gaining much attention in the medical image analysis field for the clinical treatment and follow up study. Traditional sparse representation based classifiers use a random subset in a limited range. It suffers from the problem of repetition of the training samples which may prevent obtaining optimal subset having all variations. To overcome this problem, nested cross-validation based adaptive sparse representation algorithm is newly proposed. The novelty of the work are: (i) a novel strategy for optimal subset selection, (ii) adaptively selects an optimal subset, (iii) ability to overcome the problems like overfitting, underfitting and bias results, (iv) better accuracy in all variations of training samples, and (v) newly applied to pathological brain classification problem. The proposed system is based on a hybrid methodology of feature selection followed by classification. The gray level co-occurrence matrix is used to extract the spatial texture feature vectors of the brain MRI samples. The nested cross-validation based adaptive sparse representation algorithm is used for classification. It uses a nested cross-validation technique to obtain the optimal value of the subset size (N) based on maximum classification accuracy. The results demonstrate the superiority of the proposed algorithm over the state-of-the-art methods. (C) 2018 Elsevier Ltd. All rights reserved.
Keywords:
Pathological brain classification
Gray level co-occurrence matrix
Nested cross-validation based adaptive sparse representation algorithm
Nested cross-validation technique
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
7.5
Papers:
2.9W
Citations:
10.2W