基于矩阵分解模型的图片数据库构建

Construction of picture database based on matrix decomposition model

现有的图片数据库存在图片处理效率低的缺陷,为了解决上述问题,提出基于矩阵分解模型的图片数据库构建方法。采用中值滤波法与迭代法预处理采集的多来源图片,以此为基础,求取图片颜色分量平均值与方差作为特征参数,以提取的图片特征参数为依据,通过触发器编写检索算法代码,得到图片特征索引结果。经过正态分布算法归一化图片数据,通过矩阵分解模型搭建数据库逻辑结构,为商户推荐用户偏好的图片,实现了基于矩阵分解模型图片数据库的构建。通过实验结果表明,与现有的图片数据库相比,构建的图片数据库极大的提升了图片处理效率,充分说明构建的基于矩阵分解模型的图片数据库具备更好的使用性能。

The existing image database has the defect of low efficiency of image processing.In order to solve the above problems,a method of image database construction based on matrix decomposition model is proposed.The median filter method and iterative method are used to preprocess the collected multisource images.On this basis,the average value and variance of the color components of the images are calculated as the feature parameters.Based on the extracted image feature parameters,the retrieval algorithm code is written by trigger,and the image feature indexing results are obtained.After normalizing the image data by normal distribution algorithm,the logical structure of the database is built by matrix decomposition model,and the image database based on matrix decomposition model is constructed by recommending the user preferred images for merchants.The experimental results show that compared with the existing image database,the constructed image database greatly improves the efficiency of image processing,which fully shows that the constructed image database based on matrix decomposition model has better performance.