Image Watermarking Menggunakan Metode Spektrum Tersebar Dalam Domain Transformasi Discrete Cosine Transform (dct) Dan Compressive Sensing Sebagai Serangan

Abstract

Abstrak Teknologi biomedis sekarang ini berupa data digital, salah satu contohnya yaitu data citra medis digital pasien. Data citra digital sangat mudah diakses dan disebarluaskan, namun rentan untuk dimanipulasi. Karena itu diperlukan suatu teknik guna memberikan keamanan pada data citra medis digital pasien. Dalam penelitian ini telah dilakukan perancangan sistem menggunakan teknik watermarking pada aplikasi Matlab.Teknik watermarking adalah teknik penyisipan tanda air dalam domain transformasi citra. Transformasi citra yang digunakan yaitu Discrete Cosine Transform (DCT) dengan metode spread spectrum. Spread spectrum merupakan metode penyebaran sinyal informasi yang berguna untuk memperkuat citra transformasi. Teknik watermarking juga tidak luput dari jenis serangan citra yaitu compressive sensing. Jenis serangan ini sama seperti serangan kompresi lainnya, di mana citra terwatermark dikompresi berdasarkan measurement rate yang ditentukan. Berdasarkan hasil analisis perancangan sistem rata-rata nilai Peak Signal to Noise Ratio (PSNR) dari 5 resolusi citra ter-watermark yaitu di atas 40 dB. Nilai parameter Peak Signal to Noise Ratio (PSNR) rekonstruksi citra ter-watermark serangan compressive sensing paling baik pada resolusi citra 1024×1024 yaitu 46,33 dB dan paling buruk pada resolusi 128×128 yaitu 19,38 dB. Hasil analisis pengujian ekstraksi citra watermark tanpa serangan compressive sensing yaitu semua resolusi citra mendapatkan nilai Bit Error Rate (BER) 0 dan Structural Similarity Index Matrix (SSIM) 1. Sedangkan menggunakan serangan compressive sensing hasil ekstraksi watermark yang paling baik pada resolusi 1024×1024 yaitu 0, untuk yang paling buruk pada resolusi 128×128 yaitu 0,46. Kata Kunci: watermarking, spread spectrum, compressive sensing, DCT Abstract Biomedical technology is now in the form of digital data, one example is the patient's digital medical image data. Digital image data is very easy to access and disseminate, but is vulnerable to manipulation. Because of that, a technique is needed to provide security for the patient's digital medical image data.In this research system design has been done using watermarking techniques in the Matlab application. Watermarking technique is the technique of inserting a watermark in the domain of image transformation. Image transformation used is Discrete Cosine Transform (DCT) with spread spectrum method. Spread spectrum is a method of spreading information signals that is useful for strengthening transformation images. Watermarking techniques also do not escape from the type of image attack that is compressive sensing. This type of attack is the same as other compression attacks, where the watermarked image is compressed based on the specified measurement rate. Based on the analysis of system design the average value of Peak Signal to Noise Ratio (PSNR) of 5 watermarked image resolutions is 40 dB. Peak Signal to Noise Ratio (PSNR) parameter reconstruction of the image of watermarked compressive sensing attacks is best at 1024×1024 image resolution which is 46.33 dB and worst at 128×128 resolution is 19.38 dB.The results analysis of watermark image extraction testing without compressive sensing attacks that all image resolutions get the value of Bit Error Rate (BER) 0 and Structural Similarity Index Matrix (SSIM) 1 and by using the compressive sensing attack the best watermark extraction results at 1024×1024 resolution are 0, for the worst resolution at 128×128 which is 0.46. Keywords: watermarking, spread spectrum, compressive sensing, DCT