Open conchincradle opened 1 year ago
Inspired by visual perception theories,
first-order statistics of both its color and its responses to multiscale and multi-orientation filters, namely the steerable pyramid.
starting from a white noise image, histogram matchings are performed to the image alternately in the image domain and the steerable pyra- mid domain, so that the corresponding output histograms match the ones of the input texture
the aim of texture synthesis algorithms visually similar to and pixel-wise different from the input texture
The output texture is initialized with a white noise image. Histogram matchings are performed to this output texture image alternately in the image domain and a multiscale transform domain, namely the steerable pyramid, until all the output histograms match the ones of the input texture.
Given an input image, it is obtained by first splitting the image into a high frequency part and a low frequency part and then by sequentially applying bandpass oriented filters to the low frequency image followed by downsampling. This results in a sequence of images having different sizes, referred to as a pyramid, each corresponding to a certain scale and orientation, apart from the high frequency and the low frequency residuals.
纹理合成再探 - 一种基于金字塔的方法 重点在于用一系列特定选择的一阶线性滤波器来捕捉原始纹理中的空间特征
the number of pyramid scales is the more influential parameter: using the highest value of this parameter permits to take into account all scales of the texture. Besides, using a large number of scales enables to eliminate the low frequency blotchy artifact
Histogram Matching
The Heeger-Bergen Pyramid-Based Texture Synthesis Algorithm
The Heeger-Bergen Pyramid-Based Texture Synthesis Algorithm Thibaud Briand et.al 2014