A long term learning method in CBIR systems by defining semantic templates
Esmat Rashedi, Hossein Nezamabadi–pour · 2012
This paper provides a long term learning method in CBIR systems by defining a kind of semantic template to represent semantic concepts of images. Information, which is achieved in the short learning technique, is used in construction of semantic templates. A similarity function is introduced to find the similarity between queries and semantic templates. The experimental results confirm the effectiveness of the proposed method.