Text-to-Image Generation Using Deep Learning

The Text-to-Image generation tasks has stand out as an innovation in the field of computer vision (CV). Recent innovations and advancements in generative models have led to the development of various text-to-image generation techniques. This paper presents a comprehensive implementation of a text to image generation system leveraging Stable Diffusion, a diffusion based generative model that can produce high-quality images with fine-grained details. The system is integrated with a Flask-based web application, providing users with a user-friendly interface to generate images from textual prompts. The project is designed to operate efficiently on CPUs, making it accessible in resource constrained environments. This implementation serves as a bridge for individuals and organizations seeking to harness generative AI technologies without the need of expensive hardware setups like high-end devices and GPUs. Its accessibility and simplicity make it more suitable for small-scale creative needs and inspire innovation.