Paper Title
Enhancing Personalized Sticker Image Generation Using Stable Diffusion: A Comparative Study With Gans and State-of-The-Art Models
Abstract
Stable Diffusion emerges as a promising alternative to Generative Adversarial Networks (GANs) for image
generation, offering improved stability and quality. This study demonstrates Stable Diffusion's effectiveness in generating
high-quality sticker images from textual cues, expanding its application beyond traditional image synthesis. Compared to
StackGAN and other leading models like Stable Diffusion XL, OpenDalleV1.1, OpenJourney, and Latent Consistency
Model (LCM) LoRA, Stable Diffusion excels in image quality, stability, and versatility. The study highlights Stable
Diffusion's potential to revolutionize personalized sticker image generation, establishing it as a leading technology in image
synthesis
Keywords - Stable Diffusion, Generative Adversarial Networks, Image Generation