Modelling of an Intelligent and Secured Image Retrieval Model by Employing Deep Belief Network

Searching out a particular image from huge image repositories is a very complicated and time-consuming task. Many image searching techniques and algorithms have been designed but they are not secure and are not capable to uphold the smart devices. So, to overcome all these issues, a novel image retrieval system has been designed in this paper which is both intelligent and encrypted and also a web application is developed of the framed model. The deep features are extracted by using a neural architecture i.e., Deep belief networks (DBN). To retrieve the better relevant images and that too with less retrieval time, clustering technique is applied on the features. The model is named as DBN clustering encrypting model. Web application of this model has been developed which support smart devices to enter and to retrieve the images. The precision rate of the DBN clustering encrypting model is high and retrieval time of this model is less as compared with the various other models. The designed system has been implemented on two benchmark datasets i.e., Wang and Coil-100 and the performance is compared with many related Content based image retrieval (CBIR) techniques.