摘要
针对现有医学图像检索方法中,检索算法速度较慢、准确性较低以及远程检索时占用大量带宽等问题,设计并实现了一种基于移动Agent技术的医学图像检索方法,重点讨论了该方法实现的关键技术,包括提升框架下的M带整数小波变换和提取多通道纹理特征的检索算法.模拟实验结果表明,该方法实时性较强,占用网络带宽较少,采用的算法运算速度提高近10倍、运算复杂度降低约50%,提取的纹理特征能充分表达图像内容,较好地满足了医学图像的检索要求,实现了对医学图像库快速较准确的检索.
In order to solve the existing problems in the current method of images retrieval, such as inefficient retrieval, low accuracy, and taking up too much bandwidth, a new design for medical images retrieval based on mobile Agent is proposed and implemented in this article. The key technology is discussed emphatically, including constructing M-band integer wavelet transform based on lifting framework and describing the image content by extracting the multi-channel texture feature. The experimental results show that the scheme does well in handling real-time request and occupating less bandwidth, as well as the algorithm of M-band integer wavelet transform adopted has the characteristics of higher computational efficiency, lower memory cost and wider applicability to hardware, in addition that the texture feature extracted using such an algorithm can describe the image effectively. The method is applied well and gives much better performance in medical images retrieval.
出处
《武汉大学学报（理学版）》
CAS
CSCD
北大核心
2004年第5期624-628,共5页
Journal of Wuhan University:Natural Science Edition
基金
国家自然科学基金资助项目(60132030)
