Content based Image Retrieval based on Semantics, Colour and Texture features

Progress in digital imaging techniques in the recent past fuelled by the increase in World Wide Web has escorted its way in many professional fields exploiting the opportunities to access and manipulate remotely-stored images in many propagating ways. However this has resulted in large and varied image datasels, from which the process of locating a desired image is not only difficult aod time consuming but cao be a source of frostratioo for the users. This problem of image retrieval has now been widely recognized, and thus has become an increasingly active area for research. This paper proposes a hybrid image retrieval method combining test based aod content based image retrieval techniques. The proposed method employs the use pyramid-structured wavelet transform for calculatioo of the image features aod fuzzy logic system for categorizatioo of the image into different categories for faster retrieval.