| Abstrakt: |
Signal Image separation is a significant processing task for accurate image reconstruction, which is increasingly applied to several medical imaging applications and communication areas. Most of classical separation approaches exploit frequency and time domains. These approaches, however are sensitive to noise, and thus often lead to undesirable results. In this paper, we propose a novel method of image separation. It incorporates the property of reflectance component extracted from the image and a Finite Ridgelet Transform (FRT) to obtain precise analysis of the images and thus correctly separate the images even in hardly noisy environment. We obtain the reflectance components of the target images by employing a homomorphic processing, which operates in the log domain, and thus can decompose the image into illumination and reflectance components. In addition, the homomorphic decomposition in the proposed method reduces information redundancy in the target image, and thus substantially improve the quality of image separation. We carried out extensive simulations, which demonstrate that the proposed homomorphic technique outperforms the conventional methods based on time domain and trigonometric transforms. [ABSTRACT FROM AUTHOR] |
Full text from SpringerLink