Learning Continuous Implicit Representation for Near-Periodic Patterns

TL;DR

This paper introduces a neural implicit representation method for near-periodic patterns that effectively captures global layout and local variations, improving tasks like image completion and segmentation.

Contribution

It proposes a novel coordinate-based MLP with input warping, periodicity-guided loss, and a periodicity proposal module for robust NPP modeling from a single image.

Findings

Effective on over 500 diverse images

Handles both global consistency and local variations

Improves robustness with multiple periodicity candidates

Abstract

Near-Periodic Patterns (NPP) are ubiquitous in man-made scenes and are composed of tiled motifs with appearance differences caused by lighting, defects, or design elements. A good NPP representation is useful for many applications including image completion, segmentation, and geometric remapping. But representing NPP is challenging because it needs to maintain global consistency (tiled motifs layout) while preserving local variations (appearance differences). Methods trained on general scenes using a large dataset or single-image optimization struggle to satisfy these constraints, while methods that explicitly model periodicity are not robust to periodicity detection errors. To address these challenges, we learn a neural implicit representation using a coordinate-based MLP with single image optimization. We design an input feature warping module and a periodicity-guided patch loss to handle both global consistency and local variations. To further improve the robustness, we introduce a periodicity proposal module to search and use multiple candidate periodicities in our pipeline. We demonstrate the effectiveness of our method on more than 500 images of building facades, friezes, wallpapers, ground, and Mondrian patterns on single and multi-planar scenes.