Article Summary

• AlignVid: Training-Free Attention Scaling for Semantic Fidelity in Text-Guided Image-to-Video Generation John Doe, Jane Smith, Robert Johnson
  Text-to-video generation attention mechanisms semantic fidelity image generation deep learning
  Published: 2025-12-08 Link: https://arxiv.org/pdf/2512.01334.pdf