We introduce FincGAN, a GAN framework designed to address the class imbalance in GNNs by enhancing minority sample synthesis and ensuring connectivity with sparsity-aware edge generators.