Introduction

We are witnessing the current surge in complex multimedia data. The increasing popularity of binary representation learning aims at transforming high-dimensional visual data into informative representations, and due to its compressive and discrete nature, it has been well applied in domains such as machine learning and computer science. Learning compact binary representations from images can improve the efficiency of storage, retrieval, and analytical processes. Essential aspects and applications of binary representation learning in visual imaging encompass efficient image compression to reduce storage requirements, rapid access to similar images through visual similarity searching in vast databases, and the development of content-based image retrieval systems that concentrate on the visual characteristics of images. This chapter begins by providing a formalized definition of binary representation learning and then comprehensively and specifically introduces research related to binary representation learning on visual images from four aspects: asymmetric discrete hashing, ordinal-preserving hashing, deep collaborative hashing, and trustworthy deep hashing. Additionally, we also introduce some publicly available benchmark datasets and widely used evaluation protocols that will be utilized in this book. Finally, this chapter introduces the main structure of each chapter of this book.