Summary of Mobileportrait: Real-time One-shot Neural Head Avatars on Mobile Devices, by Jianwen Jiang et al.
MobilePortrait: Real-Time One-Shot Neural Head Avatars on Mobile Devices
by Jianwen Jiang, Gaojie Lin, Zhengkun Rong, Chao Liang, Yongming Zhu, Jiaqi Yang, Tianyun Zhong
First submitted to arxiv on: 8 Jul 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI)
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 MobilePortrait method is a lightweight neural head avatars approach that integrates external knowledge to reduce learning complexity, allowing for real-time inference on mobile devices. By combining explicit and implicit keypoints for motion modeling and precomputed visual features for foreground and background synthesis, the method achieves state-of-the-art performance while reducing computational demand by over 90%. The proposed method is capable of reaching speeds of over 100 FPS on mobile devices and supports both video and audio-driven inputs. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary MobilePortrait is a new way to create realistic portraits that can be used on smartphones. It’s faster than other methods because it uses special representations of movement and images, which helps reduce the amount of information needed to learn. This makes it possible to use MobilePortrait on mobile devices in real-time. The method works well with both video and audio inputs and is very fast, running at over 100 frames per second. |
Keywords
» Artificial intelligence » Inference
