Summary of Improving Deep Generative Models on Many-to-one Image-to-image Translation, by Sagar Saxena et al.
Improving Deep Generative Models on Many-To-One Image-to-Image Translation
by Sagar Saxena, Mohammad Nayeem Teli
First submitted to arxiv on: 19 Feb 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Machine Learning (cs.LG)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary Medium Difficulty summary: This paper introduces a novel approach for improving deep generative models in image-to-image translation tasks. Building upon existing frameworks like Generative Adversarial Networks (GANs) and Diffusion Models, the authors propose an asymmetric framework that better suits many-to-one relationships between two domains. The framework is demonstrated on StarGAN V2, showcasing improved performance in both unsupervised and semi-supervised settings. To facilitate evaluation, a new benchmark dataset, Colorized MNIST, is introduced along with the Color-Recall score. This work has significant implications for applications where image-to-image translation is crucial. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Low Difficulty summary: Imagine you have pictures of different things, like animals or objects. Sometimes, these pictures are in different styles or colors, making them hard to compare. To solve this problem, scientists developed special computer models that can change one picture into another style or color. These models work well for some tasks but not for others. In this research, the authors propose a new way to make these models better at changing many pictures into one style or color. They created a new dataset of colored and black-and-white pictures to test their method. The results show that their approach works better than previous methods. |
Keywords
* Artificial intelligence * Recall * Semi supervised * Translation * Unsupervised
