رضا دیانت

Advances in electronic imaging devices ranging from high-definition televisions to displays mounted on cell phones require more advanced image enlargement techniques. An ideal technique to enlarge the image must completely reconstruct the high frequency component ‎of the image‎.‎‎ In this thesis, the comparison between different methods of down sampling and up sampling was done as the first step. Then two plans were proposed to estimate the high frequency components in the image. The first ‎scheme‎ is in wavelet domain and ‎obtain‎ using a probability model derived from relation ‎between‎ pixels in Markov tree in wavelet domain. The second plan is in the pixels domain and was done using a two-sided linear estimation ‎using ‎an ‎Auto ‎Regressive‎ ‎model.‎ Finally, simulation results were presented both ‎objective‎ and ‎subjective‎. The results showed that probability estimation was better.