Softly combining an ensemble of classifiers learned from a single convolutional neural network for scene categorization

Shuang Bai, Huadong Tang

Abstract: Highlights•An ensemble of classifiers are trained from a single convolutional neural network.•An ensemble of classifiers are softly combined for scene categorization.•Auxiliary classifiers are trained to compute dynamic weights for classifier combination.•Genetic algorithm is used to compute weights for classifier combination.