一种基于学习向量量化网络的目标提取方法

A method of object extraction based on LVQ neural network

图像目标提取是图像语义抽取及其应用的基础。为了更好地进行图像语义的抽取,提出了一种基于学习向量量化网络的目标提取方法。该方法首先将图像划分成固定大小的子块,采用改进的学习向量量化网络实现前景和背景的区分,最后提取目标。

Extraction of object in an image is a precondition for semantic image retrieval. In this article a method of object extraction based on LVQ nueual network is proposed. It first partitions an image into fixed sized blocks. Then a special method is employed to distinguish the object and background .Finally, the object,which is considered important to image retrieval is extracted.