Dose reduction using non lineal diffusion and smoothing filters in computed radiography

Abstract

The use of Computed Radiography (CR) into clinical practice has been followed by a high increase in the number of examinations performed and overdose cases in patients, especially children in pediatric applications. Computed radiographic images are corrupted by noise because either data acquisition or data transmission. The level of this inherent noise is related with the X-ray dose exposure: lower radiation exposure involves higher noise level. The main aim of this work is to reduce the noise present in a low radiation dose CR image in order to the get a CR image of the same quality as a higher radiation exposure image. In this work, we use a non lineal diffusion filtering method to reduce the noise level in a CR, this means that we are able to reduce the exposure, milliampere-second (mAs), and the dose absorbed by the patients. In order to get an optimal result, the diffusive filter is complemented with a smoothing filter with edge detection in order to preserve edges. Therefore, the proposed method consists in obtaining a good quality CR image for diagnostic purposes by selection of lower X-ray exposure jointly with a reduction of the noise. We conclude that a good solution to minimize the dose to patients, especially children in pediatric applications, in X-ray computed radiography consists in decreasing the mAs of the X-ray exposure and then processing the image with the proposed method.