Font Size: a A A

Research On Image Interpolation Method And Design Of Direction Filter Bank Based On APDCT

Posted on:2009-08-12Degree:MasterType:Thesis
Country:ChinaCandidate:F HeFull Text:PDF
GTID:2178360272985946Subject:Signal and Information Processing
Abstract/Summary:PDF Full Text Request
With the development of digital signal processing technology, digital image processing technology develops gradually. In recent years, digital image processing technology has come into all aspects of human life, for example, in military, medicine, industry, geology and other fields.Image interpolation is a process that continuous signals are recovered through discrete digital image. In the field of digital signal processing, because of the discrete style of information store, more meaningful approach of image interpolation is to change the sampling rate of digital image, that is to say, to regenerate a higher-resolution image from the original low-resolution one. Image is a typical two-dimensional data, and the content of the image has directional property. Nowadays, more and more people do research based on image directional representation and multi-resolution image processing.This paper describes an image interpolation method based on a separable two-dimensional All Phase DCT, designs a relatively fair interpolation experiment, and compares this method with nearest neighborhood interpolation, linear interpolation, cubic B-spline interpolation, cubic convolution interpolation, wavelet interpolation with zero coefficients high-frequency subband, and edge-detected interpolation. In the meanwhile, this paper points out the unreasonable aspect of wavelet interpolation, and gives a wavelet interpolation method with gray maintained. Experimental results show that image interpolation method based on all phase DCT has a very good performance.The latter part of this paper introduces design method of all phase DCT, and designs two kinds of all phase DCT directional filter bank. One is the all phase DCT direction filter bank with rotation structure, which divides the frequency with any precision; the other has more accurate frequency division in horizontal and vertical direction, and eight-direction and sixteen-direction filter banks are implemented in this paper with this method.
Keywords/Search Tags:image interpolation, directional filter bank, discrete cosine transform, all phase DCT(APDCT)
PDF Full Text Request
Related items