Summary of Contrastive Language-image Pretrained Models Are Zero-shot Human Scanpath Predictors, by Dario Zanca et al.
Contrastive Language-Image Pretrained Models are Zero-Shot Human Scanpath Predictors
by Dario Zanca, Andrea Zugarini, Simon Dietz, Thomas R. Altstidl, Mark A. Turban Ndjeuha, Leo Schwinn, Bjoern Eskofier
First submitted to arxiv on: 21 May 2023
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 presents CapMIT1003, a database of captions and click-contingent image explorations collected during captioning tasks. This dataset is based on the well-known MIT1003 benchmark, offering a promising opportunity to study human attention under both free-viewing and task-driven conditions. The authors also introduce NevaClip, a novel zero-shot method for predicting visual scanpaths that combines CLIP models with biologically-inspired neural visual attention algorithms. Experimental results demonstrate that NevaClip outperforms existing unsupervised computational models of human visual attention in terms of scanpath plausibility for both captioning and free-viewing tasks. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper looks at how people pay attention to images when they’re given a task, like writing captions. They created a big dataset called CapMIT1003 that has captions and information about what people looked at while doing the task. This is cool because it lets researchers study how people’s attention works in both free-viewing (just looking) and task-driven (doing something) situations. The authors also made a new way to predict where someone will look at an image called NevaClip, which combines two different AI methods. They tested it and found that it does a better job than other ways of predicting eye movements. |
Keywords
» Artificial intelligence » Attention » Unsupervised » Zero shot
