arXiv 2603.015862026-05-08cs.CV

InterCoG：迈向空间精确图像编辑的交错链式 grounding 推理

InterCoG: Towards Spatially Precise Image Editing with Interleaved Chain-of-Grounding Reasoning

Faculty of Computing, Harbin Institute of Technology（哈尔滨工业大学计算机学院）；Zhengzhou Advanced Research Institute of Harbin Institute of Technology（哈尔滨工业大学郑州先进研究院）；Huawei Noah’s Ark Lab（华为诺亚实验室）；Artificial Intelligence Research Institute, Shenzhen University of Advanced Technology（深圳先进技术大学人工智能研究院）

Yecong Wan, Fan Li, Chunwei Wang, Hao Wu, Mingwen Shao, Wangmeng Zuo

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

本文提出InterCoG框架，通过文本与视觉的交错链式推理实现复杂场景中精细图像编辑，结合文本空间关系和视觉定位提升编辑精度。

摘要：

新兴的统一编辑模型在一般对象编辑任务中表现出色，但在复杂多实体场景中进行细粒度编辑仍具挑战，尤其是目标不显眼且需空间推理的情况。为此，我们提出InterCoG，一种新颖的文本-视觉交错链式grounding推理框架，用于复杂真实场景中的精细图像编辑。其核心思想是首先在包含空间关系细节的文本中进行物体位置推理，以明确推断编辑目标的位置和身份。然后通过生成的边界框和掩码在像素空间中进行视觉定位，最后重写编辑描述以指定预期结果。为进一步促进这一范式，我们提出了两个辅助训练模块：多模态grounding重建监督和多模态grounding推理对齐，分别强制空间定位准确性和推理可解释性。我们还构建了GroundEdit-45K数据集，包含45,000个具有详细推理注释的grounding导向编辑样本，并构建了GroundEdit-Bench用于grounding-aware编辑评估。大量实验验证了我们的方法在高度精确编辑任务中，在空间复杂且多实体场景中的优越性。

英文摘要：

Emerging unified editing models have demonstrated strong capabilities in general object editing tasks. However, it remains a significant challenge to perform fine-grained editing in complex multi-entity scenes, particularly those where targets are not visually salient and require spatial reasoning. To this end, we propose InterCoG, a novel text-vision Interleaved Chain-of-Grounding reasoning framework for fine-grained image editing in complex real-world scenes. The key insight of InterCoG is to first perform object position reasoning solely within text that includes spatial relation details to explicitly deduce the location and identity of the edited target. It then conducts visual grounding via highlighting the editing targets with generated bounding boxes and masks in pixel space, and finally rewrites the editing description to specify the intended outcomes. To further facilitate this paradigm, we propose two auxiliary training modules: multimodal grounding reconstruction supervision and multimodal grounding reasoning alignment to enforce spatial localization accuracy and reasoning interpretability, respectively. We also construct GroundEdit-45K, a dataset comprising 45K grounding-oriented editing samples with detailed reasoning annotations, and GroundEdit-Bench for grounding-aware editing evaluation. Extensive experiments substantiate the superiority of our approach in highly precise edits under spatially intricate and multi-entity scenes.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。