SIFThinker: Spatially-Aware Image Focus for Visual Reasoning

TL;DR

SIFThinker introduces a spatially-aware framework for visual reasoning that enhances attention correction and region focusing using depth-enhanced cues, leading to improved spatial understanding in multimodal models.

Contribution

It presents a novel reverse-expansion inference strategy and a reinforced training paradigm that incorporate depth-informed visual grounding for better spatial reasoning.

Findings

Outperforms state-of-the-art in spatial understanding tasks

Constructs the SIF-50K dataset for process supervision

Demonstrates strong generalization capabilities

Abstract

Current multimodal large language models (MLLMs) still face significant challenges in complex visual tasks (e.g., spatial understanding, fine-grained perception). Prior methods have tried to incorporate visual reasoning, however, they fail to leverage attention correction with spatial cues to iteratively refine their focus on prompt-relevant regions. In this paper, we introduce SIFThinker, a spatially-aware "think-with-images" framework that mimics human visual perception. Specifically, SIFThinker enables attention correcting and image region focusing by interleaving depth-enhanced bounding boxes and natural language. Our contributions are twofold: First, we introduce a reverse-expansion-forward-inference strategy that facilitates the generation of interleaved image-text chains of thought for process-level supervision, which in turn leads to the construction of the SIF-50K dataset. Besides, we propose GRPO-SIF, a reinforced training paradigm that integrates depth-informed visual grounding into a unified reasoning pipeline, teaching the model to dynamically correct and focus on prompt-relevant regions. Extensive experiments demonstrate that SIFThinker outperforms state-of-the-art methods in spatial understanding and fine-grained visual perception, while maintaining strong general capabilities, highlighting the effectiveness of our method. Code: https://github.com/zhangquanchen/SIFThinker.