| 基于图像视觉特征与LDA模型的取景图像识别 |
| |
| 引用本文： | 杨彬蔚,张怡芳.基于图像视觉特征与LDA模型的取景图像识别[J].计算机与现代化,2011(5):61-65,69. |
| |
| 作者姓名： | 杨彬蔚 张怡芳 |
| |
| 作者单位： | 浙江机电职业技术学院计算机工程学院,浙江,杭州,310053 |
| |
| 摘 要： | 对数码相机取景图像进行场景的自动识别与归类是数码相机模式自动调整的核心问题之一。本文提出一种结合图像视觉特征与LDA(Latent Dirichlet Allocation)模型的取景图像自动识别算法:在图像类别训练中,以整个训练图像类的特征码作为训练对象,其中特征码包含图像的低层视觉特征,得到各类图像的LDA模型。在图像识别中,通过计算目标图在各类图像模型中的后验概率来判断该图所属类别。实验结果较好,说明此算法可用于对数码相机取景框图像场景识别。
|
| 关 键 词： | 视觉特征 LDA模型 场景识别 训练 |
Method of Camera Scene Auto-recognizing Based on Visual Characters and LDA Model |
| |
| YANG Bin-wei,ZHANG Yi-fang.Method of Camera Scene Auto-recognizing Based on Visual Characters and LDA Model[J].Computer and Modernization,2011(5):61-65,69. |
| |
| Authors: | YANG Bin-wei ZHANG Yi-fang |
| |
| Affiliation: | YANG Bin-wei,ZHANG Yi-fang(Department of Computer Engineering,Zhejiang Institute of Mechanical and Electrical Engineering,Hangzhou 310053,China) |
| |
| Abstract: | There is a core problem about how to auto-recognize and cluster images captured by digital cameras in camera auto-adjusted mode.A new supervised learning algorithm based-on visual characters and LDA is presented,which applies kinds of image codewords which include visual characters of images as training units prepared for image categorizing.After the auto-recognition system being trained,different digital pictures can be clustered and then their matched modes will be chosen for terminal users.The test resul... |
| |
| Keywords: | visual character LDA model scene recognition training |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
