Alzheimer's disease diagnosis on structural MR images using circular harmonic functions descriptors on hippocampus and posterior cingulate cortex

Autor: Michèle Allard, Jenny Benois-Pineau, Maxim M. Mizotin, Chokri Ben Amar, Olfa Ben Ahmed, Gwénaëlle Catheline
Přispěvatelé: Laboratoire Bordelais de Recherche en Informatique (LaBRI), Université de Bordeaux (UB)-Centre National de la Recherche Scientifique (CNRS)-École Nationale Supérieure d'Électronique, Informatique et Radiocommunications de Bordeaux (ENSEIRB), Department of Computational Mathematics and Cybernetics [Moscow], Lomonosov Moscow State University (MSU), Institut de Neurosciences cognitives et intégratives d'Aquitaine (INCIA), Université Bordeaux Segalen - Bordeaux 2-Université Sciences et Technologies - Bordeaux 1-SFR Bordeaux Neurosciences-Centre National de la Recherche Scientifique (CNRS), Département de Génie Électrique de Sfax [ENIS] (CEM Lab - ENIS), École Nationale d'Ingénieurs de Sfax | National School of Engineers of Sfax (ENIS)
Rok vydání: 2015
Předmět:
Male
CBVIR
Computer science
Hippocampus
Health Informatics
Circular Harmonic Functions
Machine learning
computer.software_genre
Gyrus Cinguli
Sensitivity and Specificity
Bag-of-Visual-Words
Imaging
Three-Dimensional
Alzheimer Disease
Support Vector Machines
Cortex (anatomy)
Histogram
Image Interpretation
Computer-Assisted
medicine
Humans
Radiology
Nuclear Medicine and imaging
Visual similarity
Aged
local features
Aged
80 and over
PCA
Radiological and Ultrasound Technology
business.industry
Reproducibility of Results
Pattern recognition
Middle Aged
Image Enhancement
Magnetic Resonance Imaging
Computer Graphics and Computer-Aided Design
Support vector machine
medicine.anatomical_structure
Bag-of-words model in computer vision
[INFO.INFO-TI]Computer Science [cs]/Image Processing [eess.IV]
Posterior cingulate
Pattern recognition (psychology)
Female
Posterior Cingulate Cortex
Computer Vision and Pattern Recognition
Artificial intelligence
Mr images
business
Alzheimer’s disease
computer
Algorithms
Zdroj: Computerized Medical Imaging and Graphics
Computerized Medical Imaging and Graphics, Elsevier, 2015, pp.34. ⟨10.1016/j.compmedimag.2015.04.007⟩
ISSN: 0895-6111
DOI: 10.1016/j.compmedimag.2015.04.007
Popis: In Press; International audience; Recently, several pattern recognition methods have been proposed to automaticallydiscriminate between patients with and without Alzheimer’s disease using different imagingtechniques: sMRI, fMRI, PET and SPECT. Classical approaches in visual informationretrieval have been successfully used for analysis of structural MRI brain images. In thispaper, we use the visual indexing framework and pattern recognition analysis based onstructural MRI data to discriminate three classes of subjects: Normal Controls (NC),Mild Cognitive Impairment (MCI) and Alzheimer’s Disease (AD). The approach uses theCircular Harmonic Functions (CHFs) to extract local features from the most involvedareas in the disease: Hippocampus and Posterior Cingulate Cortex (PCC) in each slicein all three brain projections. The features are quantized using the Bag-of-Visual-Wordsapproach to build one signature by brain (subject). This yields a transformation of a full3D image of brain ROIs into a 1D signature, a histogram of quantized features. To reducethe dimensionality of the signature, we use the PCA technique. Support vector machinesclassifiers are then applied to classify groups. The experiments were conducted on a subsetof ADNI dataset and applied to the ”Bordeaux-3City” dataset. The results showed thatour approach achieves respectively for ADNI dataset and ”Bordeaux-3City” dataset; for AD vs NC classification, an accuracy of 83.77% and 78%, a specificity of 88.2% and 80.4% and a sensitivity of 78.54% and 74.7%. For NC versus MCI classification we achievedfor the ADNI datasets an accuracy of 69.45%, a specificity of 74.8% and a sensitivity of62.52%. For the most challenging classification task (AD versus MCI), we reached anaccuracy of 62.07%, a specificity of 75.15% and a sensitivity of 49.02%. The use of PCC visualfeatures description improves classification results by more than 5% compared to the useof Hippocampus features only. Our approach is automatic, less time-consuming and doesnot require the intervention of the clinician during the disease diagnosis.
Databáze: OpenAIRE
Externí odkaz: