TY - GEN
T1 - Bag of features for automatic classification of Alzheimer's disease in magnetic resonance images
AU - Rueda, Andrea
AU - Arevalo, John
AU - Cruz, Angel
AU - Romero, Eduardo
AU - González, Fabio A.
PY - 2012
Y1 - 2012
N2 - The goal of this paper is to evaluate the suitability of a bag-of-feature representation for automatic classification of Alzheimer's disease brain magnetic resonance (MR) images. The evaluated method uses a bag-of-features (BOF) to represent the MR images, which are then fed to a support vector machine, which has been trained to distinguish between normal control and Alzheimer's disease. The method was applied to a set of images from the OASIS data set. An exhaustive exploration of different BOF parameters was performed, i.e. feature extraction, dictionary construction and classification model. The experimental results show that the evaluated method reaches competitive performance in terms of accuracy, sensibility and specificity. In particular, the method based on a BOF representation outperforms the best published result in this data set improving the equal error classification rate in about 10% (0.80 to 0.95 for Group 1 and 0.71 to 0.81 for Group 2).
AB - The goal of this paper is to evaluate the suitability of a bag-of-feature representation for automatic classification of Alzheimer's disease brain magnetic resonance (MR) images. The evaluated method uses a bag-of-features (BOF) to represent the MR images, which are then fed to a support vector machine, which has been trained to distinguish between normal control and Alzheimer's disease. The method was applied to a set of images from the OASIS data set. An exhaustive exploration of different BOF parameters was performed, i.e. feature extraction, dictionary construction and classification model. The experimental results show that the evaluated method reaches competitive performance in terms of accuracy, sensibility and specificity. In particular, the method based on a BOF representation outperforms the best published result in this data set improving the equal error classification rate in about 10% (0.80 to 0.95 for Group 1 and 0.71 to 0.81 for Group 2).
UR - http://www.scopus.com/inward/record.url?scp=84865583069&partnerID=8YFLogxK
U2 - 10.1007/978-3-642-33275-3_69
DO - 10.1007/978-3-642-33275-3_69
M3 - Conference contribution
AN - SCOPUS:84865583069
SN - 9783642332746
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 559
EP - 566
BT - Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications - 17th Iberoamerican Congress, CIARP 2012, Proceedings
T2 - 17th Iberoamerican Congress on Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications, CIARP 2012
Y2 - 3 September 2012 through 6 September 2012
ER -