Summary of Adaptive Super Resolution For One-shot Talking-head Generation, by Luchuan Song et al.
Adaptive Super Resolution For One-Shot Talking-Head Generation
by Luchuan Song, Pinxin Liu, Guojun Yin, Chenliang Xu
First submitted to arxiv on: 23 Mar 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI); Image and Video Processing (eess.IV)
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 The proposed adaptive high-quality talking-head video generation method synthesizes high-resolution videos without additional pre-trained modules. This approach down-samples the one-shot source image and adaptively reconstructs high-frequency details via an encoder-decoder module, resulting in enhanced video clarity. The method consistently improves the quality of generated videos through a straightforward yet effective strategy, substantiated by quantitative and qualitative evaluations. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper shows how to create talking-head videos from just one picture of someone’s face. Usually, this requires changing pixel values or warping facial images to make the person look like they’re in different positions. However, these methods can compromise image quality. Some approaches try to improve video quality by adding extra modules, but this increases computational costs and changes the original data. This work proposes a new method that synthesizes high-quality videos without needing additional pre-trained modules. It does this by downsampling the source image and then reconstructing high-frequency details using an encoder-decoder module. |
Keywords
» Artificial intelligence » Encoder decoder » One shot
