Text-to-Image Generation
29 Citations•2022•
Yu Cheng
journal unavailable
For the last three years diffusion denoising approach with a score-based loss function became notable as well: many researches tackling problem of image generation report SOTA results.
Abstract
For the last decade we saw significant progress in Computer Vision. Image Generation is one of the direction that researchers approached. Text - guided image generation is one of the most challenging tasks in this direction. Starting 2014, using adversarial training (GANs) was predominant approach. However for the last three years diffusion denoising approach with a score-based loss function became notable as well: many researches tackling problem of image generation report SOTA results.