Early detection of liver cancer, whether from primary occurrence or from metastization is highly important to establish informed treatment decisions. Accurate delineation of the liver tissues of interest facilitates quantitative assessment of the regions of interest, treatment application, and prognosis. Segmentation of the liver in Computer Tomography (CT) images allows the extraction of the three-dimensional (3D) structure of the liver tissues in which the observation of their relative position to one another is particularly important in treatment scenarios of radiation therapy or interventional surgery planning. The adequate receptive field for the segmentation of such a big organ in CT images, from the remaining neighbouring organs was very successfully improved by the use of the state-of-the-art Convolutional Neural Networks (CNN) algorithms, however, certain issues still arise and are highly dependent of pre-or post-processing methods to refine the final segmentations. Here, the effects of Dilated Convolutional Networks is proposed, for the purpose of improving segmentation of liver tissues in CT. The introduction of a dilation module allowed the concatenation of feature maps with a richer contextual information. The hierarchical learning process given by different dilated convolutional layers is analysed quantitatively. Experiments on the MICCAI Lits challenge dataset are described achieving segmentations with a mean Dice coefficients of 95.57% and 59.36% for the liver and liver tumour, using a total number 30 CT test volumes.