FOURIER DESCRIPTORS FOR ENHANCING DISTORTED IMAGES

In today s computer networks, the amount of digital images increases rapidly and enormously. However, images may be distorted or corrupted through different types of processing, may be through the transmission channel, or been processed by end users to enhance intensity, or histogram equalization, or compression, or transformation, or even by print-and-rescan process. Thus it is important to develop an effective technique to restore the original images. In this work, a technique is proposed to solve the problem of distorted images. This technique is based on Fourier Transform and implemented with MATLAB. The basic idea of our technique is transforming the distorted image from spatial domain into Fourier or frequency domain, and then displaying the frequency response magnitude of the distortionoperator or point spread function (PSF), with defining a threshold to remove the effect of PSF, the original image is restored. As a result, an effective image enhancement technique is achieved. The feasibility and effectiveness of our technique are demonstrated by experimental results. The qualities of the enhanced images are measured and compared with some other techniques to ensure that our technique has relatively good results.