Grid artifact reduction based on homomorphic filtering in digital radiography imaging

Dong Sik Kim; Sanggyun Lee; Jung Kee Yoon

doi:10.1117/12.2006760

Translator Disclaimer

6 March 2013 Grid artifact reduction based on homomorphic filtering in digital radiography imaging

Dong Sik Kim, Sanggyun Lee, Jung Kee Yoon

Author Affiliations +

Proceedings Volume 8668, Medical Imaging 2013: Physics of Medical Imaging; 86682C (2013) https://doi.org/10.1117/12.2006760
Event: SPIE Medical Imaging, 2013, Lake Buena Vista (Orlando Area), Florida, United States

Abstract

In digital radiography imaging, the x-ray images are formed based on a multiplicative model, in which the projected image of an object is multiplied by the antiscatter grid shadow. Hence, the formed image is amplitude- modulated by the grid shadow and the resultant modulated terms appear as the grid artifacts. Since the bandwidths of the modulated terms are as wide as that of the projected image, we should employ relatively wide-bandwidth band-stop filters (BSFs) to reduce the grid artifacts. When we apply such BSFs, the object to be recovered is prone to distortion due to the wide filter bandwidth. In this paper, to reduce the signal bandwidth of the grid shadow images in reduction of the grid artifacts, applying BSFs based on the homomorphic operation is proposed by employing the logarithmic function. By taking the logarithm of the formed image, we can separate the multiplicative grid component from both projected image and the exposure of x-rays. Hence, by employing a relatively narrow and fixed-bandwidth BSFs, we can efficiently alleviate the grid artifacts independently of the strength of the grid artifacts comparing to the conventional linear approaches. For real x-ray images, the superior performance of the proposed approach is compared in this paper.

Citation Download Citation

Dong Sik Kim, Sanggyun Lee, and Jung Kee Yoon "Grid artifact reduction based on homomorphic filtering in digital radiography imaging", Proc. SPIE 8668, Medical Imaging 2013: Physics of Medical Imaging, 86682C (6 March 2013); https://doi.org/10.1117/12.2006760

ACCESS THE FULL ARTICLE