New Gaze Dataset Aims to Improve AI Explainability in Cancer Imaging

Researchers have introduced GazeXPErT, a novel dataset designed to enhance the interpretability and explainability of AI models in oncologic FDG-PET/CT scans. This dataset captures expert eye-tracking data, correlating gaze patterns with tumor detection and measurement decisions. The goal is to facilitate the development of more trustworthy and interactive AI diagnostic tools by understanding how human experts visually process medical images. Initial experiments show that gaze data can significantly improve AI model performance and predict expert intent. AI

IMPACT This dataset could lead to more trustworthy and interactive AI diagnostic tools in oncology by integrating human visual reasoning.

RANK_REASON The cluster contains an academic paper detailing a new dataset and initial experimental results. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• 3D nnU-Net
• Fluorodeoxyglucose Positron Emission Tomography/ Computed Tomography
• GazeXPErT
• Joy Wulke
• oncology
• Vision Transformers

• paper
• product
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →