Content-Based Image Retrieval: A Survey on Local and Global Features Selection, Extraction, Representation, and Evaluation Parameters
Divya Srivastava, Shashank Sheshar Singh, B. Rajitha, Madhushi Verma, Manjit Kaur, Heung-No Lee
Abstract
Consistently the web and online media producing terabytes of images. Furthermore, the picture stockpiling and important recovery is a difficult task. Content Based Image Retrieval (CBIR) is the most common way of recovering pertinent photos regarding query images from the assortment of images. CBIR profoundly relies upon three elements: determination, extraction, and portrayal of elements/features. The proposed review work discusses about these factors in detail. It begins with examining the need for CBIR and its applications. Subsequent sections discuss selecting features such as Color, Texture, Shape, and Descriptors. Feature extraction and their representation are analysed straightaway. Further, the investigation of the most recent papers and their techniques for CBIR is presented. A CBIR has also been applied using Deep Learning Techniques in recent years. An overview of these techniques is presented in this work. The proposed article is an outline of 214 papers covering everything conceivable analysis performed in the field of CBIR.