Segmentation of Liver from Abdominal CT images has received much importance due to its complexity, as the intensity of the liver is quite similar to those of other organs surrounding it. Hence, intensity based edge operators fail to segment the liver accurately. In this work, an attempt has been made to segment liver from abdominal CT images using phase-based level set method. CT images used in this work were acquired from open source online database MIDAS and nearby hospitals. These images were subjected to phase-based distance regularized level set segmentation. A distance regularization term proposed by Chumming Li et al. was used in the energy function of level set method, to maintain stability of the evolving contour. This energy function was minimized to obtain the final contour. The phase congruency measure proposed by Kovesi was used as an edge detector for the evolution of level set function. The segmentation results were compared with the ground truth using similarity measures. It was observed that the implemented technique resulted in better similarity index and they were close to the ideal values. Leakage of contour was avoided as phase information lead to robust convergence at edges. Edges and lines were found to be well distinguished using phase congruency information. As CT images are found to be helpful in detection and characterization of abnormalities in liver like hepatoma, cysts and tumors, the proposed study seems to be clinically relevant. © 2013 by ISA - instrumentation Systems & Automation Society.