Important note: On September 7, 2026, the KITTI and KITTI-360 servers had to be restored to the backed-up state of August 7, 2026. Results within this period have to be resubmitted.

Method

Text-Guided 3D Multi-Object Tracking with RegionCLIP [TG3MOT_V1]


Submitted on 19 Oct. 2026 22:31 by
Yl LIU ()

Running time:2.2 s
Environment:GPU @ 2.5 Ghz (Python)

Method Description:
TG3MOT_V1 extends YONTD-MOT with text-guided
semantic filtering. It uses RegionCLIP for 2D
detection and VoxelRCNN for 3D detection. This V1
configuration includes TSM without 3D F-EMA or GCF.
Parameters:
RegionCLIP: enabled
TSM: enabled
3D F-EMA: disabled
GCF: disabled
Car.confidence_threshold = 0.3
Car.min_frames = 1
max_ages = 48
confidence_his_max = 9
Latex Bibtex:
@article{wang2023you,
author = {Wang, Xiyang and He, Jiawei and Fu,
Chunyun and Meng, Ting and Huang, Mingguang},
title = {You Only Need Two Detectors to Achieve
Multi-Modal {3D} Multi-Object Tracking},
journal = {arXiv preprint arXiv:2304.08709},
year = {2023}
}

Detailed Results

From all 29 test sequences, our benchmark computes the HOTA tracking metrics (HOTA, DetA, AssA, DetRe, DetPr, AssRe, AssPr, LocA) [1] as well as the CLEARMOT, MT/PT/ML, identity switches, and fragmentation [2,3] metrics. The tables below show all of these metrics.


Benchmark HOTA DetA AssA DetRe DetPr AssRe AssPr LocA
CAR 77.64 % 72.84 % 83.35 % 83.23 % 79.20 % 87.78 % 89.24 % 87.62 %

Benchmark TP FP FN
CAR 32428 1964 3713

Benchmark MOTA MOTP MODA IDSW sMOTA
CAR 83.40 % 86.23 % 83.49 % 33 70.41 %

Benchmark MT rate PT rate ML rate FRAG
CAR 82.15 % 15.38 % 2.46 % 348

Benchmark # Dets # Tracks
CAR 36141 880

This table as LaTeX

This figure as: png pdf

[1] J. Luiten, A. Os̆ep, P. Dendorfer, P. Torr, A. Geiger, L. Leal-Taixé, B. Leibe: HOTA: A Higher Order Metric for Evaluating Multi-object Tracking. IJCV 2020.
[2] K. Bernardin, R. Stiefelhagen: Evaluating Multiple Object Tracking Performance: The CLEAR MOT Metrics. JIVP 2008.
[3] Y. Li, C. Huang, R. Nevatia: Learning to associate: HybridBoosted multi-target tracker for crowded scene. CVPR 2009.


eXTReMe Tracker