Optimization of search window and mask size for non-local means noise reduction algorithm in chest digital tomosynthesis: a phantom study

Abstract

Among the parameters of non-local means (NLM) noise reduction algorithm, the search window and mask size have a great influence on the quality of diagnostic medical images. In this study, we aim to optimize the NLM noise reduction algorithm in the chest digital tomosynthesis (CDT) system. The parameters of the NLM algorithm were set to a search window of 11 x 11 to 101 x 101 and a mask size of 3 x 3 to 11 x 11. The quantitative evaluation method of the acquired CDT image used coefficient of variation (COV) and contrast-to-noise ratio (CNR). COV showed improved results as the search window was increased, and CNR showed the best results at a window size of 61 x 61. In addition, we could confirm that the COV and CNR steadily improved as the size of the mask increased. In conclusion, we expect that when applying the NLM noise reduction algorithm to CDT X-ray images, appropriate parameters can be derived and applied.