An edge-preserving total variation denoising method for DECT image

Noise in CT images can be a critical issue which degrade image quality and affect diagnosis and inspection. Dual energy CT posed a promising future in better contrast and material discrimination. However, DECT images, such as electron density and atomic reconstructions, normally have their own noise features and are difficult to describe the noise property. Denoising DECT images without harm any resolution is important for practical applications. In this work, we proposed an edge-preserving total variation based denoising method by incorporating the structural information from a low-noise image available which in DECT imaging case would be low/high energy attenuation reconstruction. The method is evaluated on practical experiments and gain very good performance.