摘要
与互联网上的通用图像相比,外观设计专利图像拥有一些其独有的特点。如,外观设计专利图像中图片质量较高,噪点较少。基于以上特点,提出了一种基于分类技术的外观设计专利图像检索方法。当检索图像以较高概率被正确分类,在该类别中检索;否则,在所有图像中检索,使用索引技术来加快检索速度。
Compared to general images in Internet,appearance design patent images have some special properties.For example,the picture quality in the appearance design of the patent image is higher,and with less noise.Based on these good properties,a Content Based Design Patent Image Retrieval method with classification technology is proposed in this paper.When the query image is right classified with high probability,similar images are retrieved in this category;otherwise,similar images are retrieved in all the images.The indexing technology is used to quicken the retrieval speed.
出处
《长春工程学院学报(自然科学版)》
2017年第1期104-108,124,共6页
Journal of Changchun Institute of Technology:Natural Sciences Edition
基金
安徽省质量工程教学研究项目(2013jyxm383)