Summary of Gazesearch: Radiology Findings Search Benchmark, by Trong Thang Pham et al.
GazeSearch: Radiology Findings Search Benchmark
by Trong Thang Pham, Tien-Phat Nguyen, Yuki Ikebe, Akash Awasthi, Zhigang Deng, Carol C. Wu, Hien Nguyen, Ngan Le
First submitted to arxiv on: 8 Nov 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 Medical eye-tracking data is crucial for understanding how radiologists analyze medical images, improving the accuracy and interpretability of deep learning models. However, current datasets are dispersed, unprocessed, and ambiguous, hindering meaningful insights. To address this issue, we propose a refinement method inspired by visual search challenges, creating a curated dataset called GazeSearch for radiology findings. Each fixation sequence is purposefully aligned to locate a specific finding. We also introduce ChestSearch, a scan path prediction baseline tailored to GazeSearch. Finally, we use GazeSearch as a benchmark to evaluate the performance of state-of-the-art methods in medical imaging visual search. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Radiologists use eye-tracking data to understand how they analyze medical images. This helps make deep learning models more accurate and clear. But right now, this data is hard to work with because it’s spread out, not organized, and unclear. So, we made a new dataset called GazeSearch that has focused eye-tracking data for finding things in medical images. We also created a special tool called ChestSearch that helps predict where radiologists look when they’re searching for something. Finally, we used GazeSearch to test how well current methods work. |
Keywords
» Artificial intelligence » Deep learning » Tracking
