QuadTok: Quadtree Visual Tokenizer for Autoregressive Image Generation

Key points

• Compared to traditional approaches using 2D grids or 1D token sequences, we propose a hierarchical quadtree structure, bridging the gap between 2D spatial binding and 1D sequence-level flexibility.1
• The QuadTok tokenizer dynamically allocates representational capacity to visually intricate areas while leaving homogeneous regions at a coarse resolution.1
• Furthermore, the natural causality introduced by the tree structure seamlessly enables autoregressive image generation.1
• Additionally, leveraging the strong spatial correlation preserved by the quadtree structure, the QuadTok generator enables zero-shot spatially controlled image generation capabilities.1

Sources (1)

• [1]QuadTok: Quadtree Visual Tokenizer for Autoregressive Image Generation
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 7, 05:47 PM

  We introduce QuadTok, a novel framework for visual tokenization and autoregressive image generation.

  Compared to traditional approaches using 2D grids or 1D token sequences, we propose a hierarchical quadtree structure, bridging the gap between 2D spatial binding and 1D sequence-level flexibility.

Extractive summary: sentences quoted from the sources.