Content Based Image Retrieval Based on Feature Extraction and Classification Using Deep Learning Techniques

Abstract

The Content based image retrieval plays a dynamic role in the contemporary scenario by being utilised to extract knowledge from photos. It is also a dynamic study area for many eras and is currently being rewarded more as a result of the theatrical and increase in the volume of digital photographs. The classifications of items based on their colour, texture, pattern, shape, layout, and location within the image, as well as other factors, are indexed and categorised according to the visual content of the image. Problem is identified in extraction of features and so the challenges are overcome by deep learning techniques. Initially the classification has been carried out using retrieval-based Inception V3-NET (RIV3-NET) algorithm. The noise has to reduce and enhance displacement with the smoothness by classifying invariant data of image using enhanced deep belief networks (EDBN). The simulation results show the enhanced retrieval of image and its parametric analysis.