A Survey of Digital Image Watermarking Techniques in Spatial, Transform, and Hybrid Domains

1. Introduction

The technological development over the last two decades achieved a milestone in the digital environment due to the uprising revolution in digital media. The growth of multimedia and computer technology prominently increases the transfer of digital data which offers more advantages like better convenience and cost efficiency. On the other hand, it also creates the issue of data piracy, tampering and copyright infringement. Nowadays copying and damaging the original data which is shared on the internet is predominant because it is difficult to differentiate between the copied and the original data. Particularly the World Wide Web (WWW) transmitted information grabs the attention of pirates and hackers. Sharing data over the network will reach a large source of the community and build a concern about image privacy and security.

Steganography refers to the basic data hiding approach which aims to hide secret data and its exact recovery. It mainly focuses on embedding capacity. The data embedded image is termed as stego image. The steganography provides a covert communication between two points (transmitter and receiver). The goal of steganography is to communicate secretly by hiding some image or text within an image/audio/video. The attacker does not know the availability or existence of communication. Watermarking is a classification of steganography. Digital watermarking is the data hiding technique to secure digital information from illegal access. In recent years, digital image watermarking has become an important technological achievement for recognizing ownership information of copyright holders and ensuring multimedia security (Begum & Uddin, 2020). This method embeds the watermark data into media such as text, image, audio, or video and extracts/detects it from the watermarked information. Thus, the host medium is modified to embed the watermark data. An eavesdropper cannot remove or modify the host data because it is safeguarded by introducing watermark data.

The color image watermarking scheme as described in Figure 1 embeds the watermark image which may be binary, grayscale, or color image into a color host image by applying a certain algorithm to form the watermarked image. Based on the watermark embedding medium, the watermarking algorithm can be categorized into Text watermarking, Audio watermarking, Image watermarking and Video watermarking as shown in Figure 2. Image watermarking scheme is classified into many groups according to the embedding process, extraction process, visibility, domain, image type and resistance against attacks.

Figure 1.

Basic principle of Color Image Watermarking

The remaining sections of the paper are structured as follows. Section 2 discusses the literature review of image watermarking techniques and section 3 states the objective of the study. Sections 4 and 5 explain the Spatial and Transform domain watermarking schemes. Section 6 elaborates on the Hybrid domain watermarking schemes. Section 7 compares the performance of all the three categories of domain based watermarking schemes. Finally, Section 8 concludes the study with suggestions for future researchers to address other issues.

2. Literature Review Of Image Watermarking Techniques

The image watermarking technique embeds an additive or multiplicative watermark into the host medium. The watermark extraction or detection process (Mahto & Singh, 2021) can be categorized into Blind, Non-Blind and Semi-Blind watermarking techniques. The Blind watermarking scheme does not require the host image to extract the watermark information whereas the Non-Blind watermarking scheme requires the host image for the extraction process. In semi-blind watermarking techniques, the hidden watermark image and some other information are required instead of the host image to extract the watermark.

Figure 2.

Classification of image watermarking techniques