Non-photorealistic image rendering with a labyrinthine tiling

TL;DR

This paper introduces a novel non-photorealistic image rendering technique that creates labyrinthine tiling patterns by replacing pixel tones based on local statistical properties, enhancing visual aesthetics for specific images.

Contribution

It presents a new algorithm for non-photorealistic rendering that generates labyrinthine patterns through statistical pixel tone replacement based on local neighborhood analysis.

Findings

Produces distinctive labyrinthine tiling patterns

Enhances visual appeal for certain images

Offers a new approach to artistic image rendering

Abstract

The paper describes a new image processing for a non-photorealistic rendering. The algorithm is based on a random generation of gray tones and competing statistical requirements. The gray tone value of each pixel in the starting image is replaced selecting among randomly generated tone values, according to the statistics of nearest-neighbor and next-nearest-neighbor pixels. Two competing conditions for replacing the tone values - one position on the local mean value the other on the local variance - produce a peculiar pattern on the image. This pattern has a labyrinthine tiling aspect. For certain subjects, the pattern enhances the look of the image.