DWT Features Based Image Retrieval Using Neural Network


Author(s): Raj Kumar Sah, Moni Kishor Chakma, Pratiksha Gautam, Asif Uddin

With the Advancement of data storage and image acquisition technologies have enabled the creation of large image datasets. So, it is necessary to develop exact information systems to efficiently manage these datasets. Image classification and retrieval is one of the most services that must be supported by such systems. The most common approach used is Content Based Image Retrieval (CBIR) System. Content Based Image Retrieval (CBIR) is a technique which uses visual contents (low level features) for image indexing and retrieving rather than the metadata (keywords, tags, etc.). In this paper, a CBIR system based on Discrete Wavelet Transform (DWT) and feed Forward neural network is proposed. In indexing phase, after decomposing the training images into wavelet domain, low level features (color and texture) are extracted. Standard Deviation is used to represent color feature and Entropy value is used to represent texture feature. Feed Forward Neural Network is used to index these features of the images to insert into the database. In the retrieving phase, after decomposing the test images or query images in the DWT domain, low level features (color and texture) are extracted in the same way. Finally, feed forward neural is used to find similarity between the images from the image database using these extracted features. The proposed method is computationally less complex and faster compare to existing methods.