提升大模型对物体数量的精准计数能力，解决大数量时易出错问题。
LVLM-COUNT: Enhancing the Counting Ability of Large Vision-Language Models
- 采用分而治之策略拆解计数任务，避免重复计数。
- 在多个数据集上验证，显著提升大数量物体的计数准确率。
- 方法简单有效，适合需要高精度计数的应用场景。
计数是多种实际视觉任务中的基础操作，需同时具备物体识别与鲁棒计数能力。尽管大型视觉语言模型（LVLM）具有先进的视觉感知能力，但在计数任务中仍表现不佳。本文评估了多个LVLM在多组计数与视觉数据集上的表现，发现其在小数量物体时误差较小，但随着物体数量增加，错误显著上升。为此，我们提出一种简单有效的基线方法，通过分而治之策略增强LVLM对大规模物体的计数能力。该方法将计数问题分解为子任务，并引入机制防止物体在分割时被重复计算——这是朴素分治法的常见缺陷。我们在多个数据集与基准测试中验证了该方法的有效性，证明其可作为未来解决方案的重要评估参考。
原文摘要 · Abstract (English)
Counting is a fundamental operation for various real-world visual tasks, requiring both object recognition and robust counting capabilities. Despite their advanced visual perception, large vision-language models (LVLMs) are known to struggle with counting tasks. In this work, we evaluate the performance of several LVLMs on visual counting tasks across multiple counting and vision datasets. We observe that while their performance may be less prone to error for small numbers of objects, they exhibit significant weaknesses as the number of objects increases. To alleviate this issue, we propose a simple yet effective baseline method that enhances LVLMs' counting ability for large numbers of objects using a divide-and-conquer approach. Our method decomposes counting problems into sub-tasks. Moreover, it incorporates a mechanism to prevent objects from being split during division, which could otherwise lead to repetitive counting -- a common issue in a naive divide-and-conquer implementation. We demonstrate the effectiveness of this approach across various datasets and benchmarks, establishing it as a valuable reference for evaluating future solutions.
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。