Lithium-ion battery electrodes are composite structures that contain multiple phases including electroactive material, conductive additive and polymer binder. The stochastic microstructure thus formed consists of particulate materials and pore network. The pore scale characteristics of the electrode determine the properties of battery. An accurate estimation of microstructural characteristics is essential for structure property relation that will enable use to design and develop batteries with enhanced properties. With the advancement in X-ray micro-/nano-tomography, details of the electrode microstructure can be probed. Tomographic images of porous electrode microstructures can be obtained at various depths or from different angles. The stack of images needs to be processed for accurate segmentation of pore and particle distribution in the matrix for statistically significant microstructure reconstruction. In the present work, image-based reconstruction of Li-ion battery electrode microstructures and subsequent statistical and transport characterization is presented. In image-based microstructure reconstruction, the segmentation is a critical step that determines the transport networks in the porous matrix. Various algorithms can be employed for segmentation including watershed transform, region merging, graph partitioning. Segmentation methods can be manual, computer assisted or hybrid. In this work, a novel segmentation algorithm has been developed based on region merging method.  The segmentation parameter is determined by the fractal nature of the segments, which in the present case is the pore and active particle interface. This enables the segmentation process to be more accurate and automated. The reconstructed electrode microstructures are characterized by its fractal dimension, which is further linked to the transport properties.