Learning Hierarchical Graph Neural Networks for Image Clustering
This addresses image clustering for computer vision applications, offering a novel supervised approach that improves performance and efficiency over existing methods.
The paper tackles the problem of clustering images into an unknown number of identities using a hierarchical graph neural network (Hi-LANDER), achieving a 54% improvement in F-score and 8% increase in NMI compared to current GNN-based methods, along with a seven-fold decrease in computational cost.
We propose a hierarchical graph neural network (GNN) model that learns how to cluster a set of images into an unknown number of identities using a training set of images annotated with labels belonging to a disjoint set of identities. Our hierarchical GNN uses a novel approach to merge connected components predicted at each level of the hierarchy to form a new graph at the next level. Unlike fully unsupervised hierarchical clustering, the choice of grouping and complexity criteria stems naturally from supervision in the training set. The resulting method, Hi-LANDER, achieves an average of 54% improvement in F-score and 8% increase in Normalized Mutual Information (NMI) relative to current GNN-based clustering algorithms. Additionally, state-of-the-art GNN-based methods rely on separate models to predict linkage probabilities and node densities as intermediate steps of the clustering process. In contrast, our unified framework achieves a seven-fold decrease in computational cost. We release our training and inference code at https://github.com/dmlc/dgl/tree/master/examples/pytorch/hilander.