GENETIC-BASED APPROACH FOR SYNTHESIZING TEXTURE

檔案

本條目沒有附加檔案。

日期

2008-08-01

期刊名稱

INTERNATIONAL JOURNAL ON ARTIFICIAL INTELLIGENCE TOOLS

摘要

Intensive search schemes are commonly adopted in patched-based texture synthesis to identity the proper neighbors of each patch. However, they do not always provide perfect solutions. This work presents a genetic algorithm in evolutionary computation for patch-based texture synthesis. The representation of an input source texture is first adjusted. Then, the synthesizing result is generated and refined by the selection of individuals based on their fitness and executing crossover and mutation until the criteria are met. After the entire execution, a large output texture is generated with texture features similar to the source texture. Because this approach optimizes the genetic algorithm, the results are satisfactory.

所屬典藏

期刊論文
顯示完整詮釋資料 統計資料