Content-Based Image Retrieval Using Image Features and Database Signature Indexing

Abstract

Billions of images across the internet offer simple access to visual information and rich texture details. Social media users, business tycoons, and researchers sometimes need related pictures. The methods of searching images are very important for users and applications. Content-Based Image Retrieval (CBIR)(Ms., March 2013, Content Based Image Retrieval Algorithm Using  Colour Models) technique used to find and retrieve images from databases. In this paper, we introduced a CBIR technique that uses texture, color, and morphological features of images as input queries in order to search related images from a particular dataset. The image features are stored separately in a feature database in the form of a matrix. The proposed system is not only fast but efficient enough and takes less storage. The databases are categorized and indexed with stored images’ signatures to get more accuracy and efficiency. Histogram intersection and Euclidean distance are used to compare the distance of YCBCR color features. Our proposed method accumulates a recall rate of less than 0.041, which is an exceptional result of the approach.