A Hybrid Technique for Image Searching Using Content Based Image Retrieval Based on YCbCr Color Transformation

Abstract

Identification or Searching of any digital information from existing database archives on the internet has an important part of daily routine life. A major part of this growing body of information is digital images and videos. Retrieval System based on Content has been a essential part of current research problems with usage in handling the gigabytes of unlabeled image data that are generated and stored digitally in huge repositories, containing information of visuals on web or computer systems. These images stored as unstructured multimedia data in different databases spread all over the world.  To overcome this shortcoming and in trying to incorporate certain amount of visual perception characteristic of YCbCr transformation technique in Content based image retrieval system is commonly applied along with the feature vector design in transformation is proposed in the work with a low complexity Histogram design matching algorithm is used. The Simulated results show the proposed technique has effective in the current existing search algorithms. With obtained results the average accuracy of searching algorithm is coming out to be 95% applied on the Wang Dataset of image data base.