Image Multidiffusion Algorithms for AI Generative Art
Ayush Chauhan, Rahul Singh Chauhan, Ankita Nainwal, Anamika Arora, Chandradeep Bhatt · 2023
Artificial intelligence (AI) research is becoming a growing trend as a result of recent improvements in machine learning. Development and grasping knowledge of art is perhaps the most subtle field of interest in the ongoing study of the “Human Vs AI” relationship. Several fascinating projects have grown up at the intersection of AI and art, but understanding and appreciating the beauty of art is still thought to be only human capacities. This paper discusses pertinent features and discoveries made by several scholars in the creation of AI art. Many scholars use GAN, but as this field of study continues to advance, many new techniques have emerged, including DALL-E, DALL-E 2, stable diffusion, multi-diffusion, DIFF EDIT, SEMSTYLE, LAFITE, and Mirror GAN. This demonstrates that there are numerous techniques for creating images, but the primary difficulty is deciding which technique is appropriate and used for text to image synthesis. So, we talked about some of the key techniques and their methodology. Three crucial techniques-DALL-E, stable diffusion, and GAN-that serve as the foundation for the creation of artificially intelligent art are discussed. Additionally, we talked about the way they work and fundamental design. Furthermore, there is also a section of the literature review that discusses the work, methodology, and restrictions of 8 researchers' models. These techniques significantly increase public interest in this subject. This paper concludes with highlights of some attractive application of the main theme in the current world.