An Effective Content Based Image Retrieval System Using Deep Learning Based Inception Model

At present times, the rapid generation of digital images has resulted in a requirement to improve the searching and retrieving process of images from huge databases. The major difficulty is the way of retrieving relevant images from massive databases with minimal time and maximum accuracy. In this view, this paper presents a new content-based image retrieval (CBIR) model using the Deep Learning-based Inception v3 Model called DLIM to effectively retrieve the images from the databases. The DLIM model will extract the features of all the images that exist in the database and store them as a feature vector. Once the query image (QI) is provided, the DLIM model will extract the features from the QI and determine the similarity measurement with the features present in the database. The images with the highest likeness in features are retrieved from the database. The performance of the DLIM model has been validated using a set of images from the Corel 10 K database. The simulation results demonstrated the extraordinary retrieval performance of the DLIM technique in terms of recall and precision.