摘要
如何跨越低层视觉特征和高层语义特征的鸿沟 ,是目前计算机视觉的一个难点。文中首先在低层视觉特征上提取图像的主要颜色 ,然后利用语义网络建立低层视觉特征和高层语义特征之间的关联 ,最后 ,为了提高检索效率 ,把相关反馈引进到图像检索系统中。实验证明 ,该方法可以应用到大量的数据库上 ,取得了较好的检索准确率。
How to narrow the gap between low level features and h i gh level features is still a challenging problem in computer vision field. The w ord-net is used to solve this problem and applied it into image retrieval. Fir stly, dominant colors from the images are extracted. Then the links between the dominant colors and image′s content using the word-net are established. And i n the semantics based image retrieval system, relevance feedback is used to impr ove the retrieval accuracy. Experimental results show that the algorithm can be applied to the large amount of images and the high accuracy is obtained.
出处
《南京航空航天大学学报》
EI
CAS
CSCD
北大核心
2005年第1期75-78,共4页
Journal of Nanjing University of Aeronautics & Astronautics
关键词
主颜色特征
语义网络
图像语义
相关反馈
dominant colors
word-net
image retrieval
relevance feed back
