Semi-automatic 3D construction of liver using single view CT images

Liver is the largest and most important organ within the human body. It is important to detect if any abnormality lies in the liver and aid surgeons in their planning before performing the surgery. In our work, we segment the liver from CT images using the distance regularized edge based level set method and demonstrate the application of the developed algorithm in performing 3D construction of liver using only the axial slices from the abdominal CT scan. The 3D construction of liver is achieved by stacking the evolved contours of individual slices over one another. The initialization is done by evolving the initial input contour to the liver boundaries. The above resulting contour then behaves as the initial contour for the adjacent slices, thereby making this process a semi-automatic one. By controlling certain parameters such as the number of iterations, standard deviation and window size of the Gaussian blurring kernel an optimal segmentation result can be generated and no interfacing with third party toolkit is required.