节点文献
基于纹理特征与BP神经网络的一类图像检索
A Kind of Image Retrieval Method Based on Texture Features and BP Neural Networks
【摘要】 <正> 1 引言随着网络通信及多媒体技术的发展,特别是因特网的广泛应用,图像作为一种越来越重要的信息载体得到了广泛的应用。融合图像理解技术,直接针对静止图像或视频帧的图像特征进行处理,在高度信息化的今天,已成为内容图像库中图像信息组织和管理不可
【Abstract】 In view of texture-based image retrieval,a kind of image retrieval method is proposed in this paper which extracts and selects texture features based on the co-occurrence matrix and then classifies them with BP Neural Networks. Experiments results show that this method is very effective for retrieving a series of images of numbers in vehicle license plates.
【关键词】 Texture-based image retrieval;
the Co-occurrence matrix;
Texture analysis;
BP neural net-works;
【Key words】 Texture-based image retrieval; the Co-occurrence matrix; Texture analysis; BP neural net-works;
【Key words】 Texture-based image retrieval; the Co-occurrence matrix; Texture analysis; BP neural net-works;
【基金】 国家自然科学基金69772026
- 【文献出处】 计算机科学 ,Computer Science , 编辑部邮箱 ,2000年06期
- 【分类号】TP391.3
- 【被引频次】16
- 【下载频次】200