Interaction-Consistent Object Removal via MLLM-Based Reasoning

2026-02-01

AI Summary

This work addresses the semantic inconsistency in existing image editing methods that neglect interaction traces when removing target objects. It formally introduces the problem of interaction-consistent object removal and proposes REORM, a novel framework that leverages a multimodal large language model (MLLM) to reason about interaction-related elements requiring simultaneous removal. The framework integrates mask-guided editing with a self-correction mechanism to achieve semantically coherent results. To facilitate evaluation, the authors construct ICOREval, the first benchmark dedicated to this task, on which REORM significantly outperforms existing approaches. The results validate the effectiveness of the proposed modular reasoning architecture and its optimization for local deployment.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Large Multimodal Models (LMMs)
• Knowledge Representation and Reasoning: Common-Sense Reasoning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Image-based object removal often erases only the named target, leaving behind interaction evidence that renders the result semantically inconsistent. We formalize this problem as Interaction-Consistent Object Removal (ICOR), which requires removing not only the target object but also associated interaction elements, such as lighting-dependent effects, physically connected objects, targetproduced elements, and contextually linked objects. To address this task, we propose Reasoning-Enhanced Object Removal with MLLM (REORM), a reasoningenhanced object removal framework that leverages multimodal large language models to infer which elements must be jointly removed. REORM features a modular design that integrates MLLM-driven analysis, mask-guided removal, and a self-correction mechanism, along with a local-deployment variant that supports accurate editing under limited resources. To support evaluation, we introduce ICOREval, a benchmark consisting of instruction-driven removals with rich interaction dependencies. On ICOREval, REORM outperforms state-of-the-art image editing systems, demonstrating its effectiveness in producing interactionconsistent results.

Problem

• Interaction-Consistent Object Removal
• Semantic Consistency
• Image Editing
• Object Removal
• Interaction Evidence

Innovation

• Interaction-Consistent Object Removal
• Multimodal Large Language Model
• Reasoning-Enhanced Editing
• Self-Correction Mechanism
• ICOREval