Due to the high complexity of geometry-deterministic wireless channel modeling and the difficulty in its implementation, geometry-based stochastic channel modeling (GSCM) approaches have been used to evaluate wireless systems. This paper introduces a new method to model any GSCM by training a generative neural network based on images formed by channel parameters. Although generative neural networks are known to capture complicated data distributions, training with the raw data of channel parameters corresponding to a specific environment yields the increased complexity of the implementation. To overcome this issue, we process the channel parameters in the form of images and utilize them to train a generative model, which substantially reduces the complexity of implementation and training time. Furthermore, through a case study, we show that the generative model trained with our proposed data-to-image mapping method faithfully represents the distributions of the original data under general wireless conditions.