期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2019
卷号：97
期号：22
页码：3273-3282
出版社：Journal of Theoretical and Applied
摘要：In this paper, properties of signals, such as speech and image are tested after transmission over OFDM system with 16-QAM under additive white Gaussian noise (AWGN). Noise could distorts the signals. The statistical and information-theoretic properties of signals that are transmitted through OFDM system are analyzed by similarity measures to determine which of properties stands better against noise. For image, statistical similarity such as Structural Similarity Index Method (SSIM) and 2D correlation were used; also used are the information-theoretic measures such as entropy and joint histogram. On other hand, Pearson Correlation Coefficient (PCC), Tanimoto coefficient and Mel-frequency cepstral coefficients (MFCC) were used for speech signals. Mean Squared Error (MSE) was also used as a similarity measure for both single- and dimensional signals. The coefficients of discrete wavelet transform (DWT) and the coefficients of the discrete cosine transform (DCT) are also tested over noisy FFT-OFDM for both image and speech signals. Results found that the MFCC-correlation measure is more stable under noise than other measures. Furthermore, DWT is more robust than DCT for both speech signal and image, where it gives higher similarity with original image under very low SNR.
