一个模型搞定计数与生成，无需特定训练即达顶尖水平
ABACUS: Adapting Unified Foundation Model for Bridging Image Count Understanding and Generation
- 用密度感知缩放和物体图定位目标，实现精准空间定位
- 通过边界感知策略减少裁剪误差，计数更准确
- 自纠错机制闭合理解与生成差距，适合多任务研究者
ABACUS 是一个统一的视觉语言模型，无需针对特定基准进行训练即可处理物体计数、人群计数、指代表达计数以及计数忠实的图像生成。该模型基于现有的 30 亿参数统一基础模型，通过三项关键创新实现对象定位：利用密度感知自适应缩放结合物体度量图实现空间定位；通过 GRPO 设计边界感知计数策略，消除裁剪边界误差；采用循环一致的 GRPO 策略，使理解分支对生成输出进行自我批评，从而在无外部标注情况下弥合理解与生成之间的差距。ABACUS 在七个基准上均取得最先进性能，超越了专用模型和更大通用模型。
原文摘要 · Abstract (English)
ABACUS is a unified vision-language model that handles object counting, crowd counting, referring-expression counting, and count-faithful image generation without any benchmark-specific training required. Our model is built on existing 3B-parameter unified foundation model and is adapted for object localization tasks using three key innovations: density-aware adaptive zooming with objectness maps for spatial grounding; a boundary-aware count policy via GRPO to eliminate crop-boundary errors; and a cycle-consistent GRPO strategy where the understanding branch self-critiques generated outputs, closing the understanding-generation gap without any external annotations. ABACUS achieves state-of-the-art results across seven benchmarks, outperforming both task-specific specialists and larger generalist models.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。