Saved in:
Bibliographic Details
Main Authors: Zhang, Zhiyuan, Li, Xiaofan, Xu, Zhihao, Peng, Wenjie, Zhou, Zijian, Shi, Miaojing, Huang, Shuangping
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2504.00379
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910900382859264
author Zhang, Zhiyuan
Li, Xiaofan
Xu, Zhihao
Peng, Wenjie
Zhou, Zijian
Shi, Miaojing
Huang, Shuangping
author_facet Zhang, Zhiyuan
Li, Xiaofan
Xu, Zhihao
Peng, Wenjie
Zhou, Zijian
Shi, Miaojing
Huang, Shuangping
contents Autonomous driving visual question answering (AD-VQA) aims to answer questions related to perception, prediction, and planning based on given driving scene images, heavily relying on the model's spatial understanding capabilities. Prior works typically express spatial information through textual representations of coordinates, resulting in semantic gaps between visual coordinate representations and textual descriptions. This oversight hinders the accurate transmission of spatial information and increases the expressive burden. To address this, we propose a novel Marker-based Prompt learning framework (MPDrive), which represents spatial coordinates by concise visual markers, ensuring linguistic expressive consistency and enhancing the accuracy of both visual perception and spatial expression in AD-VQA. Specifically, we create marker images by employing a detection expert to overlay object regions with numerical labels, converting complex textual coordinate generation into straightforward text-based visual marker predictions. Moreover, we fuse original and marker images as scene-level features and integrate them with detection priors to derive instance-level features. By combining these features, we construct dual-granularity visual prompts that stimulate the LLM's spatial perception capabilities. Extensive experiments on the DriveLM and CODA-LM datasets show that MPDrive achieves state-of-the-art performance, particularly in cases requiring sophisticated spatial understanding.
format Preprint
id arxiv_https___arxiv_org_abs_2504_00379
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle MPDrive: Improving Spatial Understanding with Marker-Based Prompt Learning for Autonomous Driving
Zhang, Zhiyuan
Li, Xiaofan
Xu, Zhihao
Peng, Wenjie
Zhou, Zijian
Shi, Miaojing
Huang, Shuangping
Computer Vision and Pattern Recognition
Autonomous driving visual question answering (AD-VQA) aims to answer questions related to perception, prediction, and planning based on given driving scene images, heavily relying on the model's spatial understanding capabilities. Prior works typically express spatial information through textual representations of coordinates, resulting in semantic gaps between visual coordinate representations and textual descriptions. This oversight hinders the accurate transmission of spatial information and increases the expressive burden. To address this, we propose a novel Marker-based Prompt learning framework (MPDrive), which represents spatial coordinates by concise visual markers, ensuring linguistic expressive consistency and enhancing the accuracy of both visual perception and spatial expression in AD-VQA. Specifically, we create marker images by employing a detection expert to overlay object regions with numerical labels, converting complex textual coordinate generation into straightforward text-based visual marker predictions. Moreover, we fuse original and marker images as scene-level features and integrate them with detection priors to derive instance-level features. By combining these features, we construct dual-granularity visual prompts that stimulate the LLM's spatial perception capabilities. Extensive experiments on the DriveLM and CODA-LM datasets show that MPDrive achieves state-of-the-art performance, particularly in cases requiring sophisticated spatial understanding.
title MPDrive: Improving Spatial Understanding with Marker-Based Prompt Learning for Autonomous Driving
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2504.00379