Unsupervised Scale-Invariant Multispectral Shape Matching

Alignment between non-rigid stretchable structures is one of the hardest tasks in computer vision, as the invariant properties are hard to define on one hand, and on the other hand no labelled data exists for real datasets. We present unsupervised neural network architecture based upon the spectrum of scale-invariant geometry. We build ontop the functional maps architecture, but show that learning local features, as done until now, is not enough once the isometric assumption breaks but can be solved using scale-invariant geometry. Our method is agnostic to local-scale deformations and shows superior performance for matching shapes from different domains when compared to existing spectral state-of-the-art solutions.