Content Based Image Retrieval Using Color and Texture Features

Images are widely used nowadays and image retrieval as one of the interesting applications. It is an appropriate case to be implemented in CPU using new algorithms. In a collection of digital images people can efficiently make use of image retrieval tools to retrieve the images using different methods for image retrieval. Image processing technique like Image Retrieval which is a computationally intensive task can be made to exploit in CPU to extract information more efficiently. Image Retrieval finds wide range of application especially in Art Collections, e.g. Fine Arts Museum of San Francisco, Medical Image Databases, e.g. CT, MRI, Ultrasound, The Visible Human, Scientific Databases, e.g. Earth Sciences, and in General Image Collections for Licensing, e.g. Corbis, Getty Images. In defense, these techniques are used Color, Texture, Shape and others. Existing systems are primitive in terms of optimal utilization of CPU using previous algorithms; implementing image retrieval on CPU using new algorithms or other platforms provides greater speedup and scalability. The major Task of this paper is to implement Color Features using (Average Intensity, Average Colors and Color Histogram) and Texture Feature Extraction using ( Hough Lines, SIFT, Blob Detection, Gradient Orientation and Magnitude and Contour Detection) in CPU. These features are used to search images in an image database which is similar like image query. Retrieval system is evaluated by using Recall and Precision and Average precision measures.