Self Supervised Clustering of Traffic Scenes using Graph Representations
This work addresses the challenge of grouping traffic scenes for applications like autonomous driving, but it is incremental as it builds on existing graph and self-supervised techniques.
The paper tackles the problem of clustering traffic scenes without manual labels by using a self-supervised method based on graph embeddings and a Siamese network, achieving clusters with common semantic characteristics as evaluated on the INTERACTION dataset.
Examining graphs for similarity is a well-known challenge, but one that is mandatory for grouping graphs together. We present a data-driven method to cluster traffic scenes that is self-supervised, i.e. without manual labelling. We leverage the semantic scene graph model to create a generic graph embedding of the traffic scene, which is then mapped to a low-dimensional embedding space using a Siamese network, in which clustering is performed. In the training process of our novel approach, we augment existing traffic scenes in the Cartesian space to generate positive similarity samples. This allows us to overcome the challenge of reconstructing a graph and at the same time obtain a representation to describe the similarity of traffic scenes. We could show, that the resulting clusters possess common semantic characteristics. The approach was evaluated on the INTERACTION dataset.