Today, the internet has been widely used in many areas of business and people's lives, Accordingly, a higher requirement to the traditional retrieval system is proposed for processing a significant number of image files. How to find the required image information quickly and accurately, and also avoid human label,low efficiency,being unable to Auto-process, which are the issues traditional retrieval system has, that all make content-based image retrieval systems become a hot research.The image features are classified into two parts which are high-level semantic features and low-level visual features including colour, texture, shape, spatial relationship, etc. Because of the limits of the related subject and field such as computer intelligence, low-level visual features are primary in the content-based image retrieval system at present.In this research, application fields of the content-based image retrieval will be introduced in the round, and the methods of extracting and describing low-level visual features including colour, texture, shape, space, will be discussed. Finally, a colour-based image retrieval system will be developed by the author's experience in master period. |