A Feed-Forward Neural Network and Euclidean Distance based Approach to CBIR using Image Queries

A content based image retrieval system is developed as an efficient, accurate and fast image retrieval system, where the user will give an image as input to get the similar images as output. The similar images are fetched from the class to which the query image belongs. Similar images are decided on the basis of the visual similarity of query image to the retrieved images. The visual features are extracted using the color and texture features of the images present in the database. These features are used to train the feed forward neural network which decides on the class of the query image during testing. The distance between query image and database images of the identified class are calculated. Among these images, 10 images with least distance values are selected and displayed as output. This application will reduce time and retrieval complexity as compared to other existing approaches or applications.