Research On Main Texture Extraction Algorithm Based On Adaptive Color Component

In today's information age,visualization technology,virtual reality technology and reality enhancement technology developing fast and popularize.These technologies rely on a large number of realistic models to meet people's visual needs,and high-quality textures can greatly improve the reality of the model,so the society has an urgent need for high quality textures.However,the color distribution of texture is random,and the existing extraction methods are relatively complex in terms of algorithm complexity and time overhead,which makes texture measurement is very difficult and quality does not meet requirements.Aiming at these problems,we propose a main texture extraction algorithm based on color and structure information under multi-resolution,which can greatly improve the texture extraction speed,and the extracted main texture and the corresponding synthesis result are very impressive.After extracting the high-quality texture,it can greatly improve the precision and efficiency of the subsequent processing steps,such as synthesizing large texture images that conform to human vision,and provide more realistic 3D models for virtual reality and augmented reality field.This paper extracts the main texture region,not a particular texture feature,because of the diversity of texel,none of the textures can be extracted by a single method,and through the study found that the general image of the main texture area,the main texture area refers to a region with a large number of uniform texture elements.Using the main texture region,we can obtain a class of main texture regions with a large number of texture elements,more in line with human vision,and then use this area to synthesize new texture regions.Synthetic texture is more realistic.At the same time,the versatility of this algorithm has also greatly improved.The concept of main texture based on Lu's paper.In this paper,the main texture extraction algorithm based on color and structure information is realized under multi-resolution.In this algorithm,by analyzing the color histogram of each resolution of color components can best distinguish the main texture and impurities,extraction and the main texture component on each resolution,combined with the structure information of texture of the final vote to get the final main texture region.We use multiresolution analysis of the texture area,to ensure that the extracted main texture is stable,is consistent with the visual texture of the region.At the same time,by selecting the color component by selecting the color component,it is necessary to solve the problem of extracting the texture with the different color components.The extraction efficiency of the texture is improved,the quality of the texture extraction is improved,which provides a solution for real-time extraction and synthesis of texture extraction.In order to verify the validity of the algorithm,the author conducted a comparative experiment on MSRA-1000 data set,experiments show that this algorithm can extract the texture region of high quality,each part of the algorithm and the computation time complexity is relatively low.Therefore,this algorithm can be applied to real-time extraction of textures.