Summary of A Lightweight Spatiotemporal Network For Online Eye Tracking with Event Camera, by Yan Ru Pei et al.
A Lightweight Spatiotemporal Network for Online Eye Tracking with Event Camera
by Yan Ru Pei, Sasskia Brüers, Sébastien Crouzet, Douglas McLelland, Olivier Coenen
First submitted to arxiv on: 13 Apr 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 paper proposes a causal spatiotemporal convolutional network (CSCN) to efficiently process event-based data in edge computing environments. The CSCN targets low-latency inference and is designed for implementation on hardware with limited resources. It achieves this through three strategies: using a simple architecture, buffering layer outputs for online inference, and achieving high activation sparsity via regularization during training. Additionally, the paper introduces an affine augmentation strategy to alleviate dataset scarcity issues in event-based systems. The proposed model is evaluated on the AIS 2024 eye tracking challenge, reaching a score of 0.9916 p10 accuracy on the private testset. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The researchers created a new way to process data quickly and efficiently using a special type of computer hardware. They developed an algorithm that can analyze event-based data, which is important for applications like edge computing. The algorithm is designed to work well with limited resources and can even improve its performance by removing unnecessary information. To make the most of the available data, they also created a new method to add variety to the training data. The team tested their algorithm on a specific challenge and achieved impressive results. |
Keywords
* Artificial intelligence * Convolutional network * Inference * Regularization * Spatiotemporal * Tracking
