Image-based simulations at pore scale provide direct insight into the impact of the microstructure on flow and transport processes in porous media. Diffusion is an important mechanism of mass transfer in gas or liquid phases, confined in porous media. Similar to fluid flow, the diffusive transport in porous media is a strong function of pore size and structure. Although the effect of porosity, pore connectivity and constrictivity of homogeneous porous media on macroscopic properties is clear, this is not well understood for heterogeneous porous media. This study uses a dual-structural-scale medium to analyze the effect of topological and morphological parameters on effective properties such as permeability and the effective diffusivity. A synthetic porous medium was created by two sizes of small and large glass beads, and 3D pore structure image of the sample was captured by X-ray computed tomography technique. The Stokes and diffusion equations were directly solved on the extracted pore geometry of sample, using a finite element method. The results show a strong nonlinear relationship between constrictivity, as a morphological parameter, with permeability and effective diffusivity. Based on the results obtained from pore-scale imaging and modeling, the connectivity of the pore space is increased by decreasing the Euler number and consequently the permeability and effective diffusivity is increased. Good agreement between image-based computed effective diffusivity with that estimated by van Brakel and Heertjes empirical relation confirms the reliability of this relation for heterogeneous porous medium, which includes constrictivity in addition to porosity and tortuosity, as three important morphological properties.