Novel nested patch-based feature extraction model for automated Parkinson’s Disease symptom classification using MRI images

Kaplan, Ela and Altunisik, Erman and Firat, Yasemin Ekmekyapar and Barua, Prabal Datta ORCID: https://orcid.org/0000-0001-5117-8333 and Dogan, Sengul and Baygin, Mehmet and Demir, Fahrettin Burak and Tuncer, Turker and Palmer, Elizabeth and Tan, Ru-San and Yu, Ping and Soar, Jeffrey ORCID: https://orcid.org/0000-0002-4964-7556 and Fujita, Hamido and Acharya, U. Rajendra (2022) Novel nested patch-based feature extraction model for automated Parkinson’s Disease symptom classification using MRI images. Computer Methods and Programs in Biomedicine, 224:107030. pp. 1-11. ISSN 0169-2607


Abstract

Objective: Parkinson's disease (PD) is a common neurological disorder with variable clinical manifestations and magnetic resonance imaging (MRI) findings. We propose an effective, handcrafted image classification model that can accurately (i) classify different PD stages, (ii) detect comorbid dementia, and (iii) discriminate PD-related motor symptoms. Methods: Selected image datasets from three PD studies were used to develop the classification model. Our proposed novel automated system was developed in four phases: (i) texture features are extracted from the non-fixed size patches. In the feature extraction phase, a pyramid histogram-oriented gradient (PHOG) image descriptor is used. (ii) In the feature selection phase, four feature selectors: neighborhood component analysis (NCA), Chi2, minimum redundancy maximum relevancy (mRMR), and ReliefF are used to generate four feature vectors. (iii) Two classifiers: k-nearest neighbor (kNN) and support vector machine (SVM) are used in the classification step. A ten-fold cross-validation technique is used to validate the results. (iv) Eight predicted vectors are generated using four selected feature vectors and two classifiers. Finally, iterative majority voting (IMV) is used to attain general classification results. Therefore, this model is named nested patch-PHOG-multiple feature selectors and multiple classifiers-IMV (NP-PHOG-MFSMCIMV). Results: Our presented NP-PHOG-MFSMCIMV model achieved 99.22%, 98.70%, and 99.53% accuracies for the collected PD stages, PD dementia, and PD symptoms classification datasets, respectively. Significance: The obtained accuracies (over 98% for all states) demonstrated the performance of developed NP-PHOG-MFSMCIMV model in automated PD state classification.


Statistics for USQ ePrint 49976
Statistics for this ePrint Item
Item Type: Article (Commonwealth Reporting Category C)
Refereed: Yes
Item Status: Live Archive
Additional Information: Files associated with this item cannot be displayed due to copyright restrictions.
Faculty/School / Institute/Centre: Current - Faculty of Business, Education, Law and Arts - School of Business (18 Jan 2021 -)
Faculty/School / Institute/Centre: Current - Faculty of Business, Education, Law and Arts - School of Business (18 Jan 2021 -)
Date Deposited: 26 Jul 2022 04:16
Last Modified: 10 Oct 2022 01:26
Uncontrolled Keywords: PD image classification; Nested patch division; Local binary pattern; Local phase quantization; Neighborhood component analysis; Image classification
Fields of Research (2020): 46 INFORMATION AND COMPUTING SCIENCES > 4609 Information systems > 460902 Decision support and group support systems
Socio-Economic Objectives (2020): 20 HEALTH > 2001 Clinical health > 200101 Diagnosis of human diseases and conditions
Identification Number or DOI: https://doi.org/10.1016/j.cmpb.2022.107030
URI: http://eprints.usq.edu.au/id/eprint/49976

Actions (login required)

View Item Archive Repository Staff Only