Block3D: Efficient Text-to-3D Generation via Block-Wise Diffusion

  • 类型:arxiv
  • 标识:2608.19567
  • 链接:https://arxiv.org/abs/2608.19567
  • 主分类:multimodal
  • 形态:method
  • 被引:1
  • 被引来源:Semantic Scholar
  • S2被引:1
  • OpenAlex被引:0
  • 影响力被引:0
  • TLDR:Block3D is proposed, a block-wise diffusion framework that partitions the discrete shape-token sequence into contiguous blocks, generates the blocks autoregressively, and jointly denoises all tokens within the current block and introduces confidence-guided intra-block correction, which revises low-confidence tokens before each block is finalized.
  • OpenAlex ID:W7203935880
  • OpenAlex DOI:10.48550/arxiv.2608.19567
  • DOI:10.48550/arxiv.2608.19567
  • DOI来源:OpenAlex
  • 开放获取:green
  • 开放获取链接:https://arxiv.org/pdf/2608.19567
  • OpenAlex更新:2026-08-31
  • 待LLM分类:否
  • 标题中文:Block3D:通过块级扩散实现高效文本到 3D 生成
  • TLDR中文:本文提出 Block3D,一种块级扩散框架,将离散形状 token 序列划分为连续块,自回归地生成各块,并联合去噪当前块内的所有 token,同时引入置信度引导的块内修正机制,在每块定稿前对低置信度 token 进行修订。
  • 来源文件:
  • /inbox/tom/_candidates/2026-08-25-agent-rag-longcontext-candidates.json
  • [S2 enrich]
  • [OpenAlex backfill]