Encoding pyramids by Labeling RAAM

摘要

In this paper we present preliminary results on the application of Labeling RAAM (LRAAM) for encoding pyramids. The LRAAM is a neural network able to develop reduced representations of labeled directed graphs. The capability of such representations to preserve structural similarities is demonstrated on a pyramid. We suggest to exploit this skill in data compression and/or to discover scale affine auto-similarities.