Saved in:
Bibliographic Details
Main Authors: Cheng, Yanchun, Wang, Rundong, Yang, Xulei, Prakash, Alok, Rus, Daniela, Ang Jr, Marcelo H, Li, ShiJie
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.06985
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915842124414976
author Cheng, Yanchun
Wang, Rundong
Yang, Xulei
Prakash, Alok
Rus, Daniela
Ang Jr, Marcelo H
Li, ShiJie
author_facet Cheng, Yanchun
Wang, Rundong
Yang, Xulei
Prakash, Alok
Rus, Daniela
Ang Jr, Marcelo H
Li, ShiJie
contents Spatial reasoning from monocular images is essential for autonomous driving, yet current Vision-Language Models (VLMs) still struggle with fine-grained geometric perception, particularly under large scale variation and ambiguous object appearance. We propose a simple yet effective perception-aware multimodal reasoning framework that equips VLMs with explicit object-centric grounding ability. Instead of relying on textual bounding-box outputs, each referred object is represented using all Visual Reference Tokens (VRTs) within its spatial extent, enabling visual evidence and textual reasoning to be processed jointly in a unified token space. To further strengthen cross-modal interaction, we construct a Multimodal Chain-of-Thought (MM-CoT) dataset that injects aligned visual and textual reasoning signals. A deterministic ordering strategy is introduced to make supervision over inherently unordered VRT sets fully compatible with the VLM's autoregressive next-token prediction. With only standard supervised fine-tuning, our method achieves substantial improvements on the SURDS benchmark, outperforming previous approaches - including those using RL-based post-training - by a large margin across both single-object and multi-object tasks. These results demonstrate that accurate perception and multimodal reasoning are mutually reinforcing, and together form the key to robust spatial understanding in challenging monocular driving scenarios.
format Preprint
id arxiv_https___arxiv_org_abs_2603_06985
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Perception-Aware Multimodal Spatial Reasoning from Monocular Images
Cheng, Yanchun
Wang, Rundong
Yang, Xulei
Prakash, Alok
Rus, Daniela
Ang Jr, Marcelo H
Li, ShiJie
Computer Vision and Pattern Recognition
Spatial reasoning from monocular images is essential for autonomous driving, yet current Vision-Language Models (VLMs) still struggle with fine-grained geometric perception, particularly under large scale variation and ambiguous object appearance. We propose a simple yet effective perception-aware multimodal reasoning framework that equips VLMs with explicit object-centric grounding ability. Instead of relying on textual bounding-box outputs, each referred object is represented using all Visual Reference Tokens (VRTs) within its spatial extent, enabling visual evidence and textual reasoning to be processed jointly in a unified token space. To further strengthen cross-modal interaction, we construct a Multimodal Chain-of-Thought (MM-CoT) dataset that injects aligned visual and textual reasoning signals. A deterministic ordering strategy is introduced to make supervision over inherently unordered VRT sets fully compatible with the VLM's autoregressive next-token prediction. With only standard supervised fine-tuning, our method achieves substantial improvements on the SURDS benchmark, outperforming previous approaches - including those using RL-based post-training - by a large margin across both single-object and multi-object tasks. These results demonstrate that accurate perception and multimodal reasoning are mutually reinforcing, and together form the key to robust spatial understanding in challenging monocular driving scenarios.
title Perception-Aware Multimodal Spatial Reasoning from Monocular Images
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2603.06985