摘要
为了缩减图像底层视觉特征与高层语义之间的"语义鸿沟"及减少聚类的不稳定性,论文提出了一种基于遗传算法和FCM的图像自动标注方法。该方法首先提取图像的颜色和纹理特征,然后运用遗传算法和FCM相结合的方法对图像进行聚类。最后通过支持向量机学习训练库的图像特征构造简单的多类支持向量机模型实现图像的自动标注。实验表明,该方法具有很好的图像标注性能。
In order to reduce the semantic gap between low-level visual features of image and high-level semantics and reduce the instability of cluster,a method of automatic image annotation based on genetic algorithms and FCM is proposed.Firstly,the color feature and texture feature of image are extracted.Then the method of combining genetic algorithm and FCM is used to cluster image.Finally,SVM constructs simplified multi-class support vector machine model by studying the image character in the training base to realize automatic image annotation.The experiments show that the method has good image annotation performance.
出处
《计算机与数字工程》
2015年第3期497-500,共4页
Computer & Digital Engineering
关键词
图像自动标注
遗传算法
模糊C均值聚类
简单多类支持向量机
automatic image annotation
genetic algorithm
fuzzy C-means clustering algorithm
simplified multi-class support vector machine
