An In-Field Calibration Method based on Performance-Surface for the Multiresolution Bilateral Filter Applied to X-Ray Images

Several spatial filters applied to images are available in technical and scientific literature. Besides the reduction of the noise level, some of them also aim at the preservation of the edges as well as the details. Those filters are commonly applied to the processing of X-ray medical image sequences, which are usually noised due to the low doses of radiation suitable for medical procedures but where the loss of any detail may impair a diagnosis. In this context, the Bilateral Filter is well suited. Recent papers presented the Multiresolution Bilateral Filter, which embed the Bilateral Filter in the process of Wavelet Denoising. However, the literature lacks a procedure for adjusting the parameters of the Bilateral Filter. Here, we show that an adequate calibration is required for guaranteeing increased performance. We present an in-field calibrating method for Multiresolution Bilateral Filter in order to reduce the noise level and preserve the image edges, built upon the analysis of a Performance-Surface. The proposed method surpasses the performance of the calibration methods that have already been presented in the literature.