基于分块主色法的图像无序激增数据检索仿真

针对传统图像信息检索方法存在检索效果不佳的问题,提出基于分块主色法的图像无序激增数据检索方法。方法首先按照图像内容将图像进行分块处理,并对每个分块进行HSV非均匀量化,得到图像的颜色特征;然后根据图像小块的颜色特征分布情况,对小块的颜色特征进行加权值计算;最终以小块加权的颜色值为目标特征,进行相似度估计,并计算图像之间的欧几里德距离,将距离值最小的图像视为检索结果。经过仿真验证,采用颜色特征进行图像信息检索效果较好,经过特征加权可以使图像检索更准确,且检索耗时短,由此可以说明所提方法具有较好的检索性能。

Image Unordered Data Retrieval Simulation Based on Block Primary Color Method

Due to poor retrieval effect of traditional methods,this article presented a retrieval method of image unordered increasing data based on dominant color of partition.Firstly,the image was divided into blocks based on the content of image,and then the color features of image were obtained by HSV non-uniform quantization for each block.According to the distribution of color features of small image blocks,the color features were weighted.Finally,the similarity degree was estimated by taking the small weighted color value as the target feature.Euclidean distance between images was calculated.The image with the minimum distance value was regarded as the retrieval result.Simulation shows that the proposed method has good effect in image information retrieval.Through the feature weighting,the image retrieval is more accurate and the retrieval time is short,so the proposed method has better retrieval performance.