Community detection is a widely discussed topic in network science which allows us to discover detailed information about the connections between members of a given group. Communities play a critical role in the spreading of viruses or the diffusion of information. In [1], [8] Kempe et al. proposed the Independent Cascade Model, defining a simple set of rules that describe how information spreads in an arbitrary network. In the same paper the influence maximization problem is defined. In this problem we are looking for the initial vertex set which maximizes the expected number of the infected vertices. The main objective of this paper is to further improve the efficiency of influence maximization by incorporating information on the community structure of the network into the optimization process. We present different community-based improvements for the infection maximization problem, and compare the results by running the greedy maximization method.