Image segmentation refers to the isolation of regions of interest in a digital image. In this work a novel method has been developed and validated to segment individual wood fibres in high-resolution 3D images of paper obtained using high-resolution Computed Tomography, i.e. micro CT (a video illustrating the 3D rendering of segmented fibres is provided in the on-line content). The isolation of each individual fibre in the three dimensional structure is challenging due to the porous structure of fibres and their compact arrangement. The method presented in this research segments the papermaking fibres by (i) tracking the hollow (lumen) inside the fibres, (ii) extracting the fibre walls surrounding the segmented lumen and (iii) labeling the fibres through collapsed sections by a final refinement step. Further, post processing algorithms have been developed to calculate the length and coarseness of the segmented fibres. The algorithms presented are validated on test geometries similar in complexity to the paper structure and created using hollow aluminum tubes that are bent and broken to simulate the discontinuous and tortuous nature of wood fibres. This work is a first step towards understanding the complex layered microstructure of paper and is vital to understanding its macro scale properties such as strength, opacity and permeability.