通过在低层视觉特征上提取图像的局部颜色和纹理特征,再利用模糊神经网络建立低层视觉特征和高层语义特征之间的映射,实现了图像属性信息的有效传递和可靠检索。最后,为了提高检索效率,把相关反馈引进到图像检索系统中。实验证明,该方法取得了较好的检索查全率和准确率。
Firstly color and the texture features were extracted. Secondly used the fuzzy nerve network to map low-features and the high-level semantic features. So this method had realized information of image' s attribution effective transmission and reliable retrieval. Finally, in the image retrieval system, introduced correlation feedback to enhance the retrieval efficiency. The experiment proves that it has obtained the high accuracy.