dc.description.abstract |
Deep learning involves training machines to understand patterns, classify objects, and predict outcomes. This has led to the emergence of various fields, including computer vision, natural language processing, speech recognition, robotics, healthcare, and image generation. Image generation is a fascinating task in deep learning and has many applications in areas such as art, gaming, fashion, and architecture. Image generation involves the creation of realistic images while adhering to specific constraints. Many companies have developed their own state-of-the-art image generation tools, such as DALL-E and MidJourney, which follow complex architectures to generate highly-realistic images. This thesis aims to compare the complexity of the model required to generate high-quality images with the current state-of-the-art architecture. The study will also explore the vast capabilities of image generation by developing image generation models on various datasets. By analyzing the complexity of the models required for image generation, this research can contribute to the development of more efficient and effective deep learning models for generating high-quality images. Additionally, this study can provide insights into the trade-offs between image quality and model complexity, advancing the field of image generation in deep learning. |
en_US |