An Extensive Analyis of AI - Generative Art Algorithms

• Published in: 2024 IEEE International Conference on Information Technology, Electronics and Intelligent Communication Systems (ICITEICS) pp. 1 - 6
• Main Authors: Bhatt, Harshit, Agarwal, Archit, Chauhan, Rahul, Rawat, Ruchira, Rajora, Ricky
• Format: Conference Proceeding
• Language: English
• Published: IEEE 28.06.2024
• Subjects: Artificial Intelligence; DALL E; DALL E2; DIFF EDIT; GAN; Generative adversarial networks; Machine learning; Mid journey; Sem style; Stable Diffusion; Technological innovation; Text to image; Visualization
• DOI: 10.1109/ICITEICS61368.2024.10624952

Recent developments in machine learning have significantly improved artificial intelligence (AI), which has led to the emergence of a growing trends. The fascinating area of AI and art has arisen within this wave of advancement, where the delicate balance between human creativity and AI capabilities is explored. Even though a lot of exciting projects have come to fruition at this point, the fundamental appreciation and comprehension of the inherent beauty of art still primarily constitute a component of human sensibility. The key ideas and findings made by academics in the field of AI art are examined in this paper. Although Generative Adversarial Networks (GANs) are widely used, a number of cutting-edge methods such as DALL E, DALL E-2, stable diffusion, multi-diffusion, DIFF EDIT, Midjourney and Mirror GAN have been introduced by the field's ongoing evolution. This variety of approaches demonstrates the range of image generation techniques available, making it difficult to choose the best strategy, especially when it comes to text-to-image synthesis. The paper explains the different methods used in these approaches, concentrating specifically on four core approaches: GANs, DALL E, Midjourney and Stable diffusion. It explains the basic ideas behind these methods and tell us about how they work. In addition, the paper includes a section on literature review which carefully examines the methods, results, and constraints of models developed by eight scholars in this field. This study underscores the potential for transformation within the art world and kindles a growing interest in the field of art by putting light on these techniques. The conclusion delves into captivating real-world applications of this theme, highlighting the considerable impact on contemporary art.