基于内容的可移植图像识别系统研究

A Portable Pattern Recognition System Research Based on Image Content

介绍了一种基于C/C++编程语言的图像识别系统,该系统通过直接对图像内容进行识别获取图像的完整信息。首先,对输入图像进行预处理,标准化图像内容;其次,输入训练图像集训练BP神经网络,使神经网络获得图像的各种特征;最后,使用测试图像集测试该系统的识别效果。实验表明,该系统具有较好的识别能力和较强的可移植性,是对图像进行识别的一种有效手段。

One kind of pattern recognition system is proposed, which is based on the C/C＋＋ programming language, and is used in indistinguishing the result from the image content in order to gain complete information of images. Firstly, input pictures are pretreated to form the standard input information of the neural network. Secondly, training sample is used to train the BP neural network on purpose of obtaining the image characteristic. Finally, testing sample is used in checking the system＇s recognition effect. It is proved that the system is effective in recognition and portability, which is an effective measure of image Reeognition.