Review
Open access
Jul 2026
Text-to-Image Generation via Deep Learning: A Comprehensive Review of Models, Architectures, and Future Directions
The paradigms of the generative adversarial networks (GANs), variational autoencoders (VAEs), transformer-based designs, and diffusion models, with the last one representing the state of the art in image generation models are discussed.
Abdussalam Elhanashi, Siham Essahraui, Qinghe Zheng et al.
· Applied Sciences · 0 citations