ZiQuan Wang, Jin Han, Qiang Li, HaiLing Zhou, Hui Li. "An image and legend text alignment model based on position guided enhanced attention mechanism." Scientific Reports, 2026. doi:10.1038/s41598-026-49176-w