Deep one-class probability learning for end-to-end image classification

Jia Liu, Wenhua Zhang, Fang Liu, Jingxiang Yang, Liang Xiao

Abstract: Highlights•We directly train an end-to-end classification network via one-class samples.•We model the distribution with output of the network via a novel energy function.•We design a particle swarm optimization based sampling and optimization method.•The proposed method models the discriminative information via a probabilistic model.

External IDs:doi:10.1016/j.neunet.2025.107201