Improved Content-Based Image Retrieval Technique for Query Generation in Mobile Networks

Image database searching is in rapid growth with an advancement in multimedia technology. To manage these kinds of searches Content-Based Image Retrieval is an effective tool. In this paper, existing CBIR techniques are analyzed and a new technique has been proposed which works based on Region-Based Convolutional Neural Network (RCNN). In the proposed approach first of all image dataset is uploaded to cloud and features are stored in a storage. Then Query image is enhanced, uploaded and features are extracted. After this feature set is compared with dataset and matched images are extracted and ranked as the closest match. Using this proposed methodology, the accuracy and precision values are compared and validated and it is observed that the proposed methodology shows better results than the existing techniques.