Method

TFACMOT: Dynamic 3D Multi-Object Tracking with Temporal-Fused Appearance [TFACMOT]
https://github.com/RhythmWings/TFACMOT

Submitted on 8 Jun. 2026 10:52 by
Ren xiwen (mioa)

Running time:0.01 s
Environment:1 core @ 2.5 Ghz (C/C++)

Method Description:
Most existing methods rely solely on geometric
features for association. A few
that introduce appearance features simply combine
them with geometric features
through a weighted sum, which not only slows down
the process but may even
compromise association accuracy. To address this,
we propose TFACMOT, a dynamic
3D multi-object tracking method with temporal-fused
appearance features as
constraints. Following an initial geometry-based
association, we apply an appearance
constraint to screen out incorrect matches by
evaluating the temporal-fused
appearance similarity of candidate pairs.
Additionally, we introduce a real-size filter to
eliminate obviously abnormal detections and design
an orientation-based dynamic life
cycle management module, which flexibly determines
trajectory termination by
considering both orientation and motion continuity.
Parameters:
Similarity threshold=0.75
Latex Bibtex:

Detailed Results

From all 29 test sequences, our benchmark computes the commonly used tracking metrics CLEARMOT, MT/PT/ML, identity switches, and fragmentations [1,2]. The tables below show all of these metrics.


Benchmark MOTA MOTP MODA MODP
CAR 90.64 % 86.87 % 91.01 % 89.64 %

Benchmark recall precision F1 TP FP FN FAR #objects #trajectories
CAR 93.44 % 98.62 % 95.96 % 36731 513 2578 4.61 % 41727 843

Benchmark MT PT ML IDS FRAG
CAR 85.08 % 6.77 % 8.15 % 129 182

This table as LaTeX

[1] K. Bernardin, R. Stiefelhagen: Evaluating Multiple Object Tracking Performance: The CLEAR MOT Metrics. JIVP 2008.
[2] Y. Li, C. Huang, R. Nevatia: Learning to associate: HybridBoosted multi-target tracker for crowded scene. CVPR 2009.


eXTReMe Tracker