摘要
尽管通过文本进行图像检索已经被广泛应用,但有些时候仍很难用文本来描述复杂图片的结构信息。而在基于手绘草图的图片检索中,可基于绘制草图来检索与其相关的图片,这对于用户非常有吸引力。提出一种新的边界点选择算法对边界点进行筛选和优化。通过在局部区域中对边界点进行筛选,保留了主要边界的信息,并将该方法应用于3种不同的草图检索算法中。通过在两个公开数据集上的实验,结果表明所提出的方法可同时提高检索的准确率和时间效率。
Although text-based approaches have already been used in Content-Based Image Retrieval,sometimes it is still very hard to represent an image structure precisely by keywords.Thus it would be an attractive pattern if the image user could draw a sketch and then use it to retrieve relevant images.In this paper,a novel local region-based edge point selection method is proposed and applied to three different Sketch-based Image Retrieval algorithms.The experiments on two public image datasets show that the proposed method could both increase the accuracy and efficiency of sketch-based image retrieval.
出处
《微型电脑应用》
2014年第4期34-37,共4页
Microcomputer Applications
基金
科技支撑计划项目(2012BAH59F04)
上海市科委项目(12dz1500203
12511504902)
