Search

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.

An interdisciplinary linguistic which studies the problem on language variation is so called Dialectology. The variation of language happens on the usage that is caused by the change of social environment and place for ages. Language variation can be found in the form of accent, sub dialect, dialect, or that of language. Studying a language variation means also to trace the language history. Geographically, language variation can show where the speaker comes from. Linguistics distinctions analyzed in dialectology are phonetics, morphemic, and lexical variations. However, in this study, phonological and lexical variations are discussed. Glosses used are Ogden's 850 basic words which have been translated into Bahasa. The research subjects are 18 Gayo native speakers. Six of them become primary informants and the rests are as secondary ones. Methods of acquiring data used are cakap-simak (speak and listen) as well as tulis-rekam (write and record). The discovered data from the informants are crosschecked with the standardized words of Kamus Besar Bahasa Indonesia (Indonesian Great Dictionary). Then, the distance of word variations are determined by using dialectometry formula. The result are: 165 words = 19,41% lexical differences, 305 words = 35,88% phonetically differences, and 380 words = 44,71% without difference, neutral or zero. By regarding 1% data error tolerant, it can be concluded that Gayo language is Old Malay that has “accent variation = 20,41%" compared to New Malay or Bahasa Indonesia.