What is it about?

This paper aims to extend the all-side regions of a given image. For example, given a building architecture image, our method could imagine the surrounding areas of this building and generate reasonable and vivid image. Through our method, the image could be generated smoothly with the continuous modeling capability.

Featured Image

Why is it important?

Our work develops a continuous extrapolation scheme which can extend the feature maps smoothly and seamlessly. It models the change of the semantic feature in the latent space. This scheme could help generate more realistic images with smooth and consistent outline of objects.


This paper solves the problem of image outpainting from a continuous scheme, which is interesting and useful.

penglei gao

Read the Original

This page is a summary of: Continuous Image Outpainting with Neural ODE, ACM Transactions on Multimedia Computing Communications and Applications, April 2024, ACM (Association for Computing Machinery),
DOI: 10.1145/3648367.
You can read the full text:



The following have contributed to this page