Color transfer in visual cryptography Hao Luo a , Hua Chen a , Yongheng Shang a,⇑ , Zhenfei Zhao b , Yanhua Zhang b a School of Aeronautics and Astronautics, Zhejiang University, Hangzhou 310027, China b School of Electronics and Information, Zhejiang University of Media and Communications, Hangzhou 310018, China article info Article history: Received 16 September 2013 Received in revised form 14 December 2013 Accepted 27 January 2014 Available online 4 February 2014 Keywords: (k, n) Visual cryptography model Color transfer Digital halftoning Monochrome output device Cheating prevention Cholesteric liquid crystal display abstract Visual cryptography is an important technique for image encryption. This paper proposes a color transfer scheme which can be incorporated into the (k, n) visual cryptography model. In encoder, a color image is encrypted into n noise-like binary share images. When any k or more than k shares are collected, a high quality colorful version of the secret image can be reconstructed with low complexity computations. The principle is motivated to develop a color image secret sharing for output devices such as monochrome printer or fax machines. The generated share images are still binary transparencies which can be directly produced by these low cost output devices. Meanwhile, the security of a (k, n) visual cryptography model is perfectly preserved. When stacking a qualified set of transparencies, the gray level version of secret content can be revealed by human visual system. Nevertheless, the pro- posed paradigm is cheating immune. It also can be integrated into some emerging display technologies such as cholesteric liquid crystal display. Experimental results and related examples demonstrate the effectiveness and efficiency. Ó 2014 Elsevier Ltd. All rights reserved. 1. Introduction As a powerful technique for image encryption [1–4] and secret sharing, the paradigm of visual cryptography scheme (VCS) is first introduced by Naor and Shamir [5]. In their (k, n) VCS model, a binary secret image is encrypted into n random noise-like shares (also called transparencies, shadows) and further printed on transparencies held by n participants. When superimposing a qualified set of k or more shares, the secret content is visible to human eyes. In other words, no computer participation or prior knowl- edge is required for secret decryption. Meanwhile, this VCS is perfectly secure for no secret information will be revealed when less than any k transparencies are overlaid. In recent years, many various VCS methods are proposed in literatures. Most of them can be summarized as following three aspects. An extensive survey can be referred to [6]. (1) Extend the basic VCS model for grayscale and color image encryption. In these schemes, digital halfton- ing or its produced halftone images are usually involved [7]. The transparencies could be binary or color halftone images. In order to enhance the secu- rity, some transparencies are meaningful images which are jointly produced by VCS and a given cam- ouflage image. (2) Introduce computer participation for incorporating an extra ability into the conventional VCS. These abilities include progressive transmission, confiden- tial data hiding for authentication, nearly lossless data reconstruction, etc. As a result, the applications of these VCS models are broadened in some specific scenarios. http://dx.doi.org/10.1016/j.measurement.2014.01.033 0263-2241/Ó 2014 Elsevier Ltd. All rights reserved. ⇑ Corresponding author. Address: No. 38, ZheDa Road, Yuquan Campus, Zhejiang University, School of Aeronautics and Astronautics, Hangzhou 310027, PR China. Tel.: +86 15858259064. E-mail address: luohao723@126.com (Y. Shang). Measurement 51 (2014) 81–90 Contents lists available at ScienceDirect Measurement journal homepage: www.elsevier.com/locate/measurement