敏感图像高维特征模糊测度与快速遗传检测

所谓敏感图像是指由文字及数字组成,带有明确指示意义的图像,其对于图像的分析与理解具有十分重要的意义。本文研究了敏感图像颜色,纹理及形状的高维识别特征模型及其检测分割准则,建立了敏感图像检测的单目标规划方程。提出了区域目标颜色与彩色纹理模糊测度,并将颜色与纹理信息进行融合,建立了敏感图像高维特征模糊测度函数。提出了敏感图像快速检测分割的遗传算法,设计了种群的初始化方法,选择了相应的适应度函数,提高了遗传算法的搜索效率。实验表明：所研究的理论与技术具有较好的参考价值。

Fuzzy Measurement of High Dimensional Characteristic and Fast Genetic Detection for Interesting Image Target

Interesting image,which is composed of text and numbers,has specific denoting meaning and plays very important roles in image analysis and understanding.In this paper,a high dimensional recognition characteristic model (including color,texture and shape) and detection criterion of interesting image is researched and the single object planning formulation for the detection is discussed. By integrating the fuzzy measurements of region object color and texture characteristics,the fuzzy measurement function for high dimen- sional characteristic of the interesting image is established.The genetic algorithm for fast detection of interesting image is developed and the algorithm works well by optimizing the population initialization and fitness function.The results of experiment prove the high refer- ence value of our theory and technology.