Abstract: The present invention discloses a brain image classification method based on discretized data, includes: dividing an original brain image dataset into an original training set, an original validation set, and an original test set; constructing a multi-objective function including an information loss before and after dataset discretization, a classification error rate, and a discrete data complexity, and obtaining a discretization scheme; discretizing the original training set, the original validation set and the original test set according to the discretization scheme; performing feature selection on a discrete training set and a discrete validation set, and performing feature reduction on the discrete training set, and a discrete test set using the feature selection result to obtain a reduced discrete training set and a reduced discrete test set; and training a classifier using the reduced discrete training set to classify the reduced discrete test set, to obtain a brain image data classification result.

Abstract: The present invention provides a method for extracting a neuroimaging biomarker based on an interpretable ensemble three-dimensional convolutional neural network (3DCNN) to address limitations in the prior art. The present invention derives a novel neuroimaging biomarker P-score from prediction results obtained by an ensemble three-dimensional convolutional neural network model. The solution can help researchers to conduct studies on longitudinal trajectory changes of structural magnetic resonance imaging (sMRI) during the progression of Alzheimer's disease, and analyze an association of the longitudinal trajectory changes with neurodegenerative changes of Alzheimer's disease subjects. The extracted neuroimaging biomarker can provide a basis for predicting a sequence of intervention of brain regions in the neurodegenerative changes of Alzheimer's disease patients and upcoming clinical symptoms.

Abstract: The present invention discloses a brain image classification method based on discretized data, includes: dividing an original brain image dataset into an original training set, an original validation set, and an original test set; constructing a multi-objective function including an information loss before and after dataset discretization, a classification error rate, and a discrete data complexity, and obtaining a discretization scheme; discretizing the original training set, the original validation set and the original test set according to the discretization scheme; performing feature selection on a discrete training set and a discrete validation set, and performing feature reduction on the discrete training set, and a discrete test set using the feature selection result to obtain a reduced discrete training set and a reduced discrete test set; and training a classifier using the reduced discrete training set to classify the reduced discrete test set, to obtain a brain image data classification result.

Abstract: The present invention provides a method for extracting a neuroimaging biomarker based on an interpretable ensemble three-dimensional convolutional neural network (3DCNN) to address limitations in the prior art. The present invention derives a novel neuroimaging biomarker P-score from prediction results obtained by an ensemble three-dimensional convolutional neural network model. The solution can help researchers to conduct studies on longitudinal trajectory changes of structural magnetic resonance imaging (sMRI) during the progression of Alzheimer's disease, and analyze an association of the longitudinal trajectory changes with neurodegenerative changes of Alzheimer's disease subjects. The extracted neuroimaging biomarker can provide a basis for predicting a sequence of intervention of brain regions in the neurodegenerative changes of Alzheimer's disease patients and upcoming clinical symptoms.