Method and device for recognizing repetitive texture

Abstract

The invention provides a method and a device for recognizing repetitive textures, and the method and the device belong to the technical field of image recognition. The method comprises the steps of: acquiring a first repetitive texture region and a preset scale threshold; determining a first extraction size of the first repetitive texture region based on the scale threshold; extracting a first texture block from the first repetition texture region according to the first extraction size; and identifying the first texture block. The device comprises a first acquisition module, a second acquisition module, a first determining module, a first extracting module and a recognition module. The method and the device determine the extraction size of the repetitive texture region through the preset scale threshold, and extract the texture block according to the extraction size based on the scale threshold; and a hypothesis that texture elements with non-repetitive textures have the same number of feature points is proposed, the method is applicable to scenarios where the hypothesis is tenable, is also applicable to scenarios where the hypothesis is untenable, and extends the recognition scenarios of the repetitive textures.

Description

technical field [0001] The invention relates to the technical field of image recognition, in particular to a method and device for recognizing repeated textures. Background technique [0002] Image texture is the relationship of pixels between image regions, such as the mean, variance, difference, or other more complex relationships of pixels between image regions. Repetitive textures are textures that repeat in a certain pattern within a connected region. Repetitive textures commonly exist in images, such as solid colors, grids, dots, stripes, leopard prints, houndstooth patterns, floral patterns, zebra patterns, etc. By identifying repeated textures, it can assist in identifying objects in the image, or find similar object. A texel is a fixed or very similar texture structure in a repeating texture area, and the number of texels in an area is the scale of the repeating texture. [0003] When identifying repeated textures, texture blocks are extracted from repeated textu...

AI Technical Summary

Problems solved by technology

[0006] In the above method, the texture block is extracted based on the assumption that texels of repeated textures have the same number of feature points. This assumption is not necessarily true for messy and repeated textures such as leopard prints, which limits the use scenarios of the above method.

Images