Unsupervised Dense Shape Correspondence using Heat Kernels

by   Mehmet Aygün, et al.

In this work, we propose an unsupervised method for learning dense correspondences between shapes using a recent deep functional map framework. Instead of depending on ground-truth correspondences or the computationally expensive geodesic distances, we use heat kernels. These can be computed quickly during training as the supervisor signal. Moreover, we propose a curriculum learning strategy using different heat diffusion times which provide different levels of difficulty during optimization without any sampling mechanism or hard example mining. We present the results of our method on different benchmarks which have various challenges like partiality, topological noise and different connectivity.


page 2

page 4

page 6

page 8


Schrödinger Diffusion for Shape Analysis with Texture

In recent years, quantities derived from the heat equation have become p...

Unsupervised Deep Learning for Structured Shape Matching

We present a novel method for computing correspondences across shapes us...

Statistical Translation, Heat Kernels and Expected Distances

High dimensional structured data such as text and images is often poorly...

Learning shape correspondence with anisotropic convolutional neural networks

Establishing correspondence between shapes is a fundamental problem in g...

Warp Consistency for Unsupervised Learning of Dense Correspondences

The key challenge in learning dense correspondences lies in the lack of ...

Topologically Robust 3D Shape Matching via Gradual Deflation and Inflation

Despite being vastly ignored in the literature, coping with topological ...

What Would a Graph Look Like in This Layout? A Machine Learning Approach to Large Graph Visualization

Using different methods for laying out a graph can lead to very differen...

Please sign up or login with your details

Forgot password? Click here to reset