IMAGE INTERPOLATION BASED ON INTER-SCALE DEPENDENCY IN WAVELET DOMAIN (TA-P7)

Author(s) : Dong Hun Woo (Dept. Electronic Eng. Pusan National University, South Korea) Il Kyu Eom (Dept. Information and Communicatio Eng. Miryang National University, South Korea) Yoo Shin Kim (Dept. Electronic Eng. Pusan National University, South Korea)

Abstract : Image interpolation in wavelet domain can be considered as the estimation of wavelet coefficients in the highest frequency subband. In this paper, a novel image interpolation based on inter-scale dependency in wavelet domain is proposed. In our method, Gaussian mixture model (GMM) is used to estimate the magnitude of wavelet coefficient, and the parameters of GMM are derived from subbands with no training procedure. The sign of the estimated wavelet coefficient is also obtained by using the inter-scale dependency of wavelet subbands. In simulation results, the proposed method shows improved PSNR and subjective quality compared with conventional bicubic method and the statistical method [1] that exploits hidden Markov tree (HMT) model with training procedure.

Menu