Summary of Divatrack: Diverse Bodies and Motions From Acceleration-enhanced Three-point Trackers, by Dongseok Yang et al.
DivaTrack: Diverse Bodies and Motions from Acceleration-Enhanced Three-Point Trackers
by Dongseok Yang, Jiho Kang, Lingni Ma, Joseph Greer, Yuting Ye, Sung-Hee Lee
First submitted to arxiv on: 14 Feb 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 A deep learning framework called DivaTrack is proposed to infer full-body poses from sparse three-point tracker inputs. This is challenging due to the under-constrained nature of the problem, particularly when considering diverse body sizes and activities. The approach augments the sparse inputs with linear accelerations from Inertial Measurement Units (IMUs) to improve foot contact prediction. A two-stage model conditions the lower-body pose based on foot contact and upper-body pose predictions. Additionally, the framework learns to blend predictions in two reference frames, designed for different types of motions, to stabilize the inferred full-body pose. The effectiveness of DivaTrack is demonstrated on a large dataset featuring 22 subjects performing challenging locomotion, including lunges, hula-hooping, and sitting. Real-time tracking accuracy is shown through a live demo using the Meta VR headset and Xsens IMUs. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary In this paper, scientists developed a new way to track people’s movements in digital reality. They created a system called DivaTrack that can figure out where someone’s body is, even when they’re moving around or doing different things like sitting or dancing. The old ways of tracking were limited and didn’t work well for everyone. The new system uses special sensors and computer learning to get better results. It even works in real-time, which means it can keep track of movements as they happen. This is important because it could help people interact more naturally with virtual worlds. |
Keywords
* Artificial intelligence * Deep learning * Tracking
