Summary of Cross-dataset Gaze Estimation by Evidential Inter-intra Fusion, By Shijing Wang and Yaping Huang and Jun Xie and Yi Tian and Feng Chen and Zhepeng Wang
Cross-Dataset Gaze Estimation by Evidential Inter-intra Fusion
by Shijing Wang, Yaping Huang, Jun Xie, Yi Tian, Feng Chen, Zhepeng Wang
First submitted to arxiv on: 7 Sep 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 The proposed EIF framework combines local regression with a cross-dataset branch to improve gaze estimation across diverse environments. By training datasets jointly, the model generalizes well, but mixing datasets decreases performance in original domains. The framework includes evidential regressors for uncertainty estimation and intra-evidential fusion among local regressors within each dataset. Inter-evidential fusion is achieved through a MoNIG distribution. Experimental results show notable improvements in both source and unseen domains. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Gaze prediction is important for real-world applications, but it’s tricky to get right. Most papers focus on a single dataset, but what if we combine multiple datasets? That can help the model work better in new situations. The problem is that mixing datasets makes the model worse in the original settings. To solve this issue, researchers created a new framework called EIF. It has three parts: local regression for each dataset, a cross-dataset branch to share information, and special “evidential” tools to show how sure the model is about its predictions. This approach did really well in both familiar and new environments. |
Keywords
» Artificial intelligence » Regression
