Research On Content-Based Image Retrieval

With the development of multimedia,communication and Internet .Image have a extensive application,as a vision information.Meeting a increace in a geometric series in image era, how to retrieve them effectively and rapidly from large–scale image libraries becomes an important and challenging research topic.Traditional image retrieval technology is based on text.It uses key words to note image. But it is impissible to note large-scale images.In such a embarrassedly conditon,content-based image retrieval(CBIR) technology is advanced.it is methoed to retrieve image by characteristics of the object (or region)of image,which includes color ,texture,shape and spatial relations.CBIR is a new research direction.It overcomes those disadvantages that TBIR(Text-Based Image Retrieval) holds.It has an enormous development space.In the dissertation,the author deals with the investigation.Firstly,lots of exploratory research work has been done around some key techniques of CBIR,which include low-level feature extraction,distance numeration,similarity measure,relevance feedback, and so on.some popular algorithm and techniques based color and texture have been introduced.Secondly,a Content-Based Image Retrieval system is structured by C++ programme language .In the sysem,the relevance knowledge is applied.So application value and the theory based color and texture is validated.Lastly,in order to test the veracity of the system,a typical tested image set and three experiment scheme is repared. By those experiments some conclusions is elicited.