Indonesian Journal of Electrical Engineering and Computer Science Vol. 18, No. 1, April 2020, pp. 368~376 ISSN: 2502-4752, DOI: 10.11591/ijeecs.v18.i1.pp368-376  368 Journal homepage: http://ijeecs.iaescore.com Multiple watermarking technique using optimal threshold Noruhida Alias, Ferda Ernawan Faculty of Computer Systems & Software Engineering, Universiti Malaysia Pahang, Malaysia Article Info ABSTRACT Article history: Received Aug 19, 2019 Revised Oct 21, 2019 Accepted Nov 5, 2019 The multiple watermarking technique has drawn more attention due to high demand for embedding more than one copyright. This paper proposes a multiple watermarking scheme using DWT-SVD by utilizing HVS characteristics. Our scheme embeds multiple watermarks in red and blue colors. The proposed scheme examines the coefficients of U orthogonal matrix for embedding and extracting watermarks. Watermarks are scrambled by Arnold transform before embedded in the host image in order provide additional security. The optimal thresholds for red and blue colors are revealed by finding a trade-off between normalized-cross correlation and imperceptibility from quantization steps. The experimental results demonstrate that our scheme achieves high resistant under JPEG and JPEG2000 compression for both inserted watermarks. Keywords: Arnold transforms Discrete wavelet transform HVS characteristics Multiple watermarking Singular value decomposition Copyright © 2020 Institute of Advanced Engineering and Science. All rights reserved. Corresponding Author: Ferda Ernawan, Faculty of Computer Systems & Software Engineering, Universiti Malaysia Pahang, Malaysia. Email: ferda@ump.edu.my 1. INTRODUCTION The used of the internet and transferring digital media over the network have led to distribute multimedia data. The usage of the internet allows users to copy and distribute the digital multimedia data. Digital watermarking is an alternative solution to protect the copyright of multimedia data. Digital watermarking has become popular in protecting the copyrights and authentication in the last decade [1-5]. The basic requirements of digital watermarking are its robustness, imperceptibility and security [6]. The multiple watermarking technique can protect multiple ownership of the multimedia data. It also can be used to verify the integrity of the contents. Multiple watermarking embeds more than one watermark image into the host image [7]. Due to rapid technology, multiple watermarks are needed in a variety of applications such as fingerprinting, broadcast monitoring, copyright protection and digital signatures [8-9]. This paper proposes a multiple watermarking scheme using the optimal thresholds on red and blue colors. The color host image is divided into red, green and blue colors. Red and blue color are selected to embed multiple watermarks due to less sensitive to the human visual system (HVS). The HVS entropy is utilized to determine the embedding regions. The smallest entropy value is used to select embedded blocks of the image. Selected blocks are transformed by Discrete Wavelet Transform (DWT) to obtain low-low sub band (LL), high-low sub band (HL), low-high subband (LH) and high-high subband (HH). Singular Value decomposition (SVD) is applied to LL sub band to derive U, S and V matrices. The multiple watermarks are embedded by modifying the first column of U matrix. This paper proposed the optimal thresholds for red and blue colors by examining the imperceptibility and robustness performance against the JPEG quantization step. The optimal thresholds can be achieved as a trade-off between the resistance of extracted watermark and invisible of the watermarked image.