Segmentation of Moire Fringes of Scoliotic Spines Using Filtering and Morphological Operations

For reducing uncertainties in moire pattern analysis, an accurate segmentation of moire fringes (MF) is vital. In this study, an algorithm for segmenting MFs of scoliotic spines was provided in MATLAB (R) environment by an empirically established sequence of filtering and morphological operations defined by static function parameters: (1) contrast enhancement, (2) brightness increase, (3) contrast refinement, (4) 2-D Gaussian filter, (5) dilation, (6) thresholding, (7) skeletonization. The algorithm is simple, fast to process and, for the most part of the images, follows the MFs correctly. Further research on segmenting MFs is quite promising by improving the algorithm with adaptive and dynamic solutions, and exploring ways to replace time demanding and complex image processing techniques.