| Summary: | 碩士 === 國立成功大學 === 資訊工程學系碩博士班 === 101 === In digital Image processing, noise removal is a highly demanded area of research. Impulse noise is common in images in which it arises at the time of image acquisition or transmission of images. Impulse noise can be classified into two categories, namely Fix-Valued Noise (Salt-and-Pepper Noise) and Random-Valued Impulse Noise. Most of the nonlinear filters used in removal of impulse noise work in two phases, such as detection followed by filtering only the corrupted pixels, keeping uncorrupted ones intact.
In this thesis, two efficient edge-preserving denoising techniques are proposed to re-move the fix-valued impulse noise and random-valued impulse noise from corrupted im-ages, respectively. Our method can preserve the edge feature efficiently by choosing an optimal edge in the working window, and reconstruct the detected noisy pixels with corre-sponding restored-value. Extensive experimental results demonstrate that our method can obtain better performances in terms of both subjective and objective evaluations than state-of –the-art impulse denoising techniques.
Simulation has been carried out at various noise conditions with different standard images. Comparative analysis has been made with existing standard schemes with suitable parameters such as Peak Signal to Noise Ratio (PSNR), SSIM index, fault detection and misses. It has been observed in general that the proposed scheme outperforms its counter-parts at any kinds of noise conditions.
|
|---|
