【摘要】 对海量复杂图像细小特征进行检索能够提高图像信息提取的完整性。针对当前海量复杂图像细小特征检索方法存在的检索速度慢,对计算机运行速度影响较大,且检索图像质量差问题,提出一种基于快速稀疏编码的海量复杂图像细小特征检索方法,通过欧几里得距离确定图像细小特征区域与其相邻区域的颜色差,计算图像细小特征区域与复杂图像总的颜色相似度,根据颜色相似度,对图像细小特征色调的相对频率进行计算,完成图像细小特征提取。通过引入快速稀疏编码的方法,根据提取的图像细小特征构建检索模型,根据模型求解结果,实现海量复杂图像细小特征检索。实验结果表明,所提方法图像细小特征检索的速度较快,检索对计算机运行速度的影响较小,且检索出图像的质量较高。更多还原

【Abstract】 To retrieve small feature of massive complex images can improve the completeness of image information extraction. Therefore,a method to retrieve small feature in massive complex image based on fast sparse coding was presented. Firstly,Euclidean distance was used to determine the color difference between the small feature region of image and its adjacent region. Then,the total color similarity between the small feature region and complex image was calculated. According to the color similarity,the relative frequency of tone of small feature was calculated to complete the small feature extraction. Moreover,the method of fast sparse coding was introduced and the retrieval model was built based on extracted image small features. Finally,massive complex image retrieval features was achieved based on the result of model solution. Simulation results show that the proposed method has faster retrieval in small feature of image and less influence on the running velocity. Meanwhile,the quality of retrieved image is higher.更多还原