arrow
Return

Test-Time Correction: An Online 3D Detection System via Visual Prompting

delete2025-12-09
delete0
PRE
AI
H
Hanxue Zhang
Z
Zetong Yang
Y
Yanan Sun
陈
陈力 (Li‐Qun Chen)
F
Fei Xia
F
Fatma Güney
H
Hongyang Li
DOI:10.1109/TPAMI.2025.3642076delete
deleteOriginal
deleteShare
deleteSave
Abstract

Abstract

En 中文
This paper introduces Test-time Correction (TTC), an online 3D detection system designed to rectify test-time errors using various auxiliary feedback, aiming to enhance the safety of deployed autonomous driving systems. Unlike conventional offline 3D detectors that remain fixed during inference, TTC enables immediate online error correction without retraining, allowing autonomous vehicles to adapt to new scenarios and reduce deployment risks. To achieve this, we equip existing 3D detectors with an Online Adapter (OA) module—a prompt-driven query generator for real-time correction. At the core of OA module are visual prompts: image-based descriptions of objects of interest derived from auxiliary feedback such as mismatches with 2D detections, road descriptions, or user clicks. These visual prompts, collected from risky objects during inference, are maintained in a visual prompt buffer to enable continuous correction in future frames. By leveraging this mechanism, TTC consistently detects risky objects, achieving reliable, adaptive, and versatile driving autonomy. Extensive experiments show that TTC significantly improves instant error rectification over frozen 3D detectors, even under limited labels, zero-shot settings, and adverse conditions. We hope this work inspires future research on post-deployment online rectification systems for autonomous driving.
Keywords:
3D object detection
online detection system
test-time correction

Journal

IEEE Transactions on Pattern Analysis and Machine Intelligence cover
IEEE Transactions on Pattern Analysis and Machine Intelligence
IF:
18.6
Papers:
2.1K
Citations:
9.8W

Organization

K
Koç University
Scholars:
386
Papers: 190
Citations: 6.2K
U
university of hong kong
Scholars:
3.6K
Papers: 1.7K
Citations: 0
S
Shanghai AI Lab
Scholars:
10
Papers: 8
Citations: 0
M
meituan inc.
Scholars:
1
Papers: 1
Citations: 0
G
gac
Scholars:
6
Papers: 5
Citations: 0
researcher View more organizations
Cited Papers

Cited Papers

BEVFormer: Learning Bird’s-Eye-View Representation from Multi-camera Images via Spatiotemporal Transformers
err2022-11-06
err0
PREAI
errZhiqi Li; Wenhai Wang; Hongyang Li; Enze Xie; Chonghao Sima; Tong Lu; Yu Qiao; Jifeng Dai
errShare
errSave
MonoTTA: Fully Test-Time Adaptation for Monocular 3D Object Detection
err2025-01-01
err0
PREAI
errLin,Hongbin; Zhang,Yifan; Niu,Shuaicheng; Cui,Shuguang; Li,Zhen
errShare
errSave
Monocular Quasi-Dense 3D Object Tracking
err2022-01-01
err0
errOAAI
errHou-Ning Hu; Yung-Hsu Yang; Tobias Fischer; Trevor Darrell; Fisher Yu; Min Sun
errShare
errSave
PnPNet: End-to-End Perception and Prediction With Tracking in the Loop
err2020-06-01
err0
errOAAI
errMing Liang; Bin Yang; Wenyuan Zeng; Yun Chen; Rui Hu; Sergio Casas; Raquel Urtasun
errShare
errSave
LL3DA: Visual Interactive Instruction Tuning for Omni-3D Understanding, Reasoning, and Planning
err2024-06-16
err0
PREAI
errSijin Chen; Xin Chen; Chi Zhang; Mingsheng Li; Gang Yu; Hao Fei; Hongyuan Zhu; Jiayuan Fan; Tao Chen
errShare
errSave
MonoDETR: Depth-guided Transformer for Monocular 3D Object Detection
err2023-10-01
err0
errOAAI
errRenrui Zhang; Han Qiu; Tai Wang; Ziyu Guo; Ziteng Cui; Yu Qiao; Hongsheng Li; Peng Gao
errShare
errSave
researcher View more