Optimasi Keamanan Watermarking pada Daubechies Transform Berbasis Arnold Cat Map

Abdussalam Abdussalam,De Rosal Ignatius Moses Setiadi,Christy Atika Sari,Ageng Setiyanto Noor,Eko Hari Rachmawanto

doi:10.30591/jpit.v4i1.911

Abdussalam Abdussalam, De Rosal Ignatius Moses Setiadi + Show 3 more

Open Access

https://doi.org/10.30591/jpit.v4i1.911

Copy DOI

Abstract

Digital image security using Transform Domain algorithms such as Discrete Wavelet Transform (DWT) has been widely used. To improve the security of the DWT algorithm needs to randomize the pixel coefficient, namely Arnold Cat Map (ACM). Computing ACM as one of the chaos functions is known to be fast and fits with Transform Domain. DWT has been implemented in the Daubechies filter which is the development of the Haar filer. In this paper, we proposed the message insertion model using a combination of DWT and ACM on a 512x512 piskel grayscale image and a 64x64 pixel message on the LL subband. The experiments were performed on 2 different images to determine the ability produced by the combined algroithm. The ability test for message insertion process is done through Mean Square Error (MSE), Peak Signal to Noise Ratio (PSNR) and comparation between original image histogram and image insertion histogram. While in the process of message extraction, algorithmic capability test is done by calculating Normalized Cross Correlation (NCC) and its correlation. The highest MSE result is 2.9502 and the highest PSNR is 43.4323 dB, while the NCC value is 237.3584 with correlation 0.7181.

Highlights

Kata Kunci Watermarking, Discrete Wavelet Trasform, Filter Daubechies, Arnold Cat Map, Peak Signal to Noise Ratio
In this paper, we proposed the message insertion model using a combination of Discrete Wavelet Transform (DWT) and Arnold Cat Map (ACM) on a 512x512 piskel grayscale image and a 64x64 pixel message on the LL subband
“A chaotic map based DCTSVD watermarking scheme for rightful ownership verification,” in 2014 Students Conference on Engineering and Systems, 2014, pp. 1–4

Summary

Nama Citra

Berdasarkan Tabel I, dapat disimpulkan bahwa pada seluruh citra yang digunakan untuk percobaan telah menghasilkan nilai PSNR lebih dari 40 dB. Dengan demikian hasil proses penyisipan pesan dapat diterima oleh sistem penglihatan manusia (HVS). Nilai PSNR tertinggi yang didapatkan yaitu 43.4411 dB pada citra aerial.tiff. Sedangkan 3 buah citra lainnya mendapatkan PSNR yang sama yaitu 43.4323 Db. Melalui percobaan yang telah dilakukan, diketahui bahwa format citra dengan format *.bmp dan format *.tiff tidak berpengaruh terhadap proses penyisipan pesan. Untuk citra pesan yang digunakan yaitu citra biner berukuran 64x64 piksel sesuai Gbr 3

Median Filtering

Salt and Pepper