Computer-aided volume measurement of the local homogenous region on magnetic resonance images

Diseases or some experimental factors may lead to the change of volume in a particular tissue, but it is difficult for physicians to identify these subtle changes visually. In this study, we designed and developed a computer-aided analysis software package for volume measurement of the local homogenous region on magnetic resonance (MR) images. The key algorithm integrated the threshold value method and the scan line seed fill (SLSF) algorithm. The SLSF algorithm is the combination of scan line filling and seed filling algorithm, which is one of the classic algorithms in region growing. In practice, some tissue may have complicated geometric configuration which makes the traditional segmentation methods (e.g. manual segmentation) complex or difficult to operate. Using this software, users can simply click the mouse in the region of interest, choose a seed point, then the specified tissue will be segmented and its volume will be calculated. To evaluate the software, we used it to calculate the volume of cochlear endolymph of guinea pig. The cochlea endolymph of normal guinea pig had 3810 voxels totally (when the threshold was 300), and its volume was 58.136 mm³. It proved that the software was efficient and the result was accurate. These quantitative results were convenient for physicians to do some statistical analyses and make accurate judgment or diagnosis.