Parameters Of The Image Interpolation Technology And Applications

In the field of the digital image processing, the magnifying and the resolution enhancement can be realized by image interpolation, which is very important practically in the sharpness processing and the 3-D displaying.In this thesis, an algorithm for image interpolation based on cluster analysis is proposed. The interpolation formula is not merely the weighted sum of the pixel values in the neighborhood, but determining the property of the pixel values first, then operating the interpolation according to the different cases. So, if the exact property is distinguished, high frequency information, such as edge and texture, can be held after interpolation.After an image is decomposed by wavelet, three high frequency sub-images and one low frequency sub-image are gained. In the high frequency one, the values of most pixels are close to zero, while the low frequency one is the best approach of the original image on certain scale and resolution, which contains ninety percent of the energy of the original image .Based on this fact, the second algorithm is proposed, which translates the image interpolation problem into an optimization problem.A lot of experiments in this paper validate that the two methods can achieve preferable results to some image data such as natural scenery, animal pictures and architectures. And, if combining the traditional image interpolation technique and the algorithms proposed by the thesis, better results can be obtained.