Text this: Advancing Text-to-Image Generation: A Comparative Study of StyleGAN-T and Stable Diffusion 3 under Neutrosophic Sets