# Improving Autoencoder Image Interpolation via Dynamic Optimal Transport

arXiv: https://arxiv.org/abs/2404.08900 · HTML: https://mlsanity.com/paper/2404.08900/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 13 Apr 2024.

## Built on by (most cited first, 3 of 3)

- [Joint Manifold Learning and Optimal Transport for Dynamic Imaging](https://mlsanity.com/paper/2505.11913/) (2025)
- [Neural Hamilton--Jacobi Characteristic Flows for Optimal Transport](https://mlsanity.com/paper/2510.01153/) (2025)
- [Measuring the Data](https://mlsanity.com/paper/2504.02083/) (2025)

## Abstract (arXiv)

Autoencoders are important generative models that, among others, have the ability to interpolate image sequences. However, interpolated images are usually not semantically meaningful.In this paper, motivated by dynamic optimal transport, we consider image interpolation as a mass transfer problem and propose a novel regularization term to penalize non-smooth and unrealistic changes in the interpolation result. Specifically, we define the path energy function for each path connecting the source and target images. The autoencoder is trained to generate the L2 optimal transport geodesic path when decoding a linear interpolation of their latent codes. With a simple extension, this model can handle complicated environments, such as allowing mass transfer between obstacles and unbalanced optimal transport. A key feature of the proposed method is that it is physics-driven and can generate robust and realistic interpretation results even when only very limited training data are available.