Visual Structures Helps Visual Reasoning: Addressing the Binding Problem in VLMs (2506.22146v2)

Abstract: Despite progress in Vision-LLMs (VLMs), their capacity for visual reasoning is often limited by the \textit{binding problem}: the failure to reliably associate perceptual features with their correct visual referents. This limitation underlies persistent errors in tasks such as counting, visual search, scene description, and spatial relationship understanding. A key factor is that current VLMs process visual features largely in parallel, lacking mechanisms for spatially grounded, serial attention. This paper introduces a simple yet effective intervention: augmenting visual inputs with low-level spatial structures (e.g., horizontal lines) and pairing this with a textual prompt that encourages sequential, spatially-aware parsing. We empirically demonstrate substantial performance improvements across core visual reasoning tasks. Specifically, our method improves GPT-4o visual search accuracy by 25.00%, increases counting accuracy by 26.83%, reduces edit distance error in scene description by 0.32, and enhances performance on spatial relationship tasks by 9.50% on a a 2D synthetic dataset. Furthermore, we find that the visual modification is essential for these gains; purely textual strategies, including Chain-of-Thought prompting, are insufficient and can even degrade performance. Our method enhances binding only with a single-query inference, underscoring the importance of visual input design over purely linguistically-based approaches. These findings suggest that low-level visual structuring is a powerful and underexplored direction for improving compositional visual reasoning and could serve as a general strategy for enhancing VLM performance on spatially grounded tasks.