Optimization of paintbrush rendering of images by dynamic MCMC methods

Abstract

We have developed a new stochastic image rendering method for the compression, description and segmentation of images. This paintbrush-like image transformation is based on a random searching to insert brush-strokes into a generated image at decreasing scale of brush-sizes, without predefined models or interaction. We introduced a sequential multiscale image decomposition method, based on simulated rectangular-shaped paintbrush strokes. The resulting images look like good-quality paintings with well-defined contours, at an acceptable distortion compared to the original image. The image can be described with the parameters of the consecutive paintbrush strokes, resulting in a parameter-series that can be used for compression. The painting process can be applied for image representation, segmentation and contour detection. Our original method is based on stochastic exhaustive searching which takes a long time of convergence. In this paper we propose a modified algorithm of speed up of about 2x where the faster convergence is supported by a dynamic Metropolis Hastings rule.