OCmst: One-class novelty detection using convolutional neural network and minimum spanning trees

Riccardo La Grassa, Ignazio Gallo, Nicola Landro

Abstract: Highlights•We present a novel model called (OCmst) for novelty detection problems.•Our model uses jointly CNN and graph-based model based on Minimum Spanning Tree.•We achieved the state-of-the-art in many classes of the CIFAR10 dataset.