Image Retrieval with the use of different color spaces and the texture feature
Gauri Chetan Deshpande, Megha Sunil Borse · 2011
Traditional text based methods are proven to be insufficient for retrieval of images from the large image data base. For large data base assigning the labels to each image using text is extremely time consuming and is valid for only one language at a time. Different users can assign different labels to the same image. To overcome these drawbacks images can be retrieved with the help of contents present in that image. This type of retrieval of image is called as Content Based Image Retrieval. Image contains two types of features High level features and Low level features. These features are nothing but the actual contents present in that image. Extracting these features we can retrieve the images. For low level feature color, RGB space is converted into HSV space and YCbCr space, for getting the better results. For image retrieval using texture, co-occurrence matrix is used. These low level features are used according to applications. In case of natural images color feature gives better result while for textured images co-occurrence matrix gives the better results.