Trends in AI Models for Image Generation
Trends in AI Models for Image Generation

Author(s): Bozhidar Bahov
Subject(s): Economy, Business Economy / Management, ICT Information and Communications Technologies, Socio-Economic Research
Published by: Университет за национално и световно стопанство (УНСС)
Keywords: image generation; autoencoders; convolutional neural networks; generative adversarial networks; diffusion models
Summary/Abstract: This conference paper traces the captivating evolution of image generation models, taking the reader on a journey through the transformative progress from initial concepts of Autoencoders and Convolutional Neural Networks (CNNs) to the more recent innovations of Generative Adversarial Networks (GANs) and Diffusion models. We put into perspective how this relentless march of technology, characterized by breakthroughs and improvements, has given rise to groundbreaking models like Midjourney, DALL-E 2, and Stable Diffusion, which now define the contemporary paradigm of image generation. Providing a rich, historical backdrop, this work serves as a compass to navigate the current state-of-the-art, inviting readers to reflect on the implications and potentials for future explorations in this dynamic field.

• Details
• Contents