Multiresolution Neural Networks for Imaging
This work addresses imaging tasks such as texture processing and antialiasing, but it appears incremental as an extended version of prior research.
The authors tackled the problem of multiresolution image representation by introducing MR-Net, a continuous coordinate-based neural network architecture that progressively adds finer details, resulting in a compact and efficient representation for applications like texture magnification, minification, and antialiasing.
We present MR-Net, a general architecture for multiresolution neural networks, and a framework for imaging applications based on this architecture. Our coordinate-based networks are continuous both in space and in scale as they are composed of multiple stages that progressively add finer details. Besides that, they are a compact and efficient representation. We show examples of multiresolution image representation and applications to texturemagnification, minification, and antialiasing. This document is the extended version of the paper [PNS+22]. It includes additional material that would not fit the page limitations of the conference track for publication.