Abstract: As the diversity and size of digital image collections grow exponentially, efficient image retrieval is becoming increasingly important. In general, current automatic image retrieval systems can be characterized into two categories: text-based and image content-based. For text-based image retrieval, the images are searched using the annotated text. In this framework, manual image annotation is extremely laborious and the visual content of images are difficult to be described precisely by a limited set of text terms. To overcome these difficulties, content-based image retrieval systems index images by their visual content, such as color, shape, texture, etc. It is a remarkable fact that, neither searching the images based on the content of the image nor searching the images using the annotated text may lead to an accurate result but jointly they tend to produce a perfect result; this is probably because the writers of text descriptions of images tend to leave out what is visually obvious (the color of flowers, etc.) and to mention properties that are very difficult to infer using vision (the species of the flower, say) and the content of the image depicts the description that may be left out by the writer. An efficient image retrieval system is highly desired. An algorithm which can combine both the retrieval systems i.e. Text Based and Content Based search and then filter out the common images can provide the exact solution for the underlying problem of the retrieval system. Our approach strives to implement the content based search by color and texture features of the objects present in the image using DWT, RGB color filter and color moments and text based search using the simple string match algorithm, and later using both results a similarity comparison is carried out to come up with a final result of the retrieval system. Our image extraction algorithm is based on both the content and the text based retrieval system with high recall rate.