Multiresolution Neural Networks for Imaging

• Hallison Paz IMPA
• Tiago Novello IMPA
• Vinicius Silva PUC-Rio
• Guilherme Schardong Universidade de Coimbra
• Luiz Schirmer Universidade de Coimbra
• Fabio Chagas IMPA
• Helio Lopes PUC-Rio
• Luiz Velho IMPA

Resumo

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 texture magnification, minification, and antialiasing.

Palavras-chave: Graphics, Image resolution, Neural networks, Imaging, Image representation, Multiresolution, Level fo Detail, Neural Networks, Textures