IMPLEMENTASI METODE CONTRAST LIMITED ADAPTIVE HISTOGRAM EQUALIZATION DAN LAPLACIAN OF GAUSSIAN FILTER UNTUK PENINGKATAN KONTRAS CITRA CT

Abstract

The  image  obtained  from  the  CT  scanner  has  a  relatively  low  contrast.  Image  contrast  can  be enhanced  by  using  an  algorithm  of  contrast  limited  adaptive  histogram  equalization  (CLAHE). However, the implementation of CLAHE causes an increase  of noise. Furthermore, noise can be reduced  using  a  noise  reduction  filter,  but  it  reduces  the  spatial  resolution  of  the  image.  This study proposed a combination of the CLAHE method and the Laplacian of Gaussian (LoG) filter noise reduction algorithm to obtain images with high contrast,  while maintaining the  noise and spatial  resolution.  In  this  study,  27  abdominal  CT  images  and  MTF  phantom  CT  images  were investigated using our proposed method. As a comparison, this study also evaluated the results of combination  of  CLAHE  and  Gaussian  filters.  Evaluation  of  the  results  of  image  quality improvement  was carried out qualitatively (visually) and quantitatively by calculating the mean squared  error  (MSE),  peak-signal  to  noise  ratio  (PSNR)  and  spatial  resolution  using  a modulation transfer function (MTF) of 10%. The results of this study qualitatively (visually) and quantitatively showed that the combination of CLAHE and LoG filters produces images with high contrast,  while  maintaining the  noise  and  the  spatial  resolution  of  the  image.  Compared  to  the combination of  CLAHE and Gaussian filters or CLAHE alone,  the  combination of CLAHE and Log  filters  provides less  noise  and  higher  spatial  resolution.  Thus,  combination  of  CLAHE  and LoG filters can be used as an alternative algorithm to increase the contrast in CT images.