issues
search
JonathonLuiten
/
TrackEval
HOTA (and other) evaluation metrics for Multi-Object Tracking (MOT).
MIT License
975
stars
243
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
how to work trackeval with yolo and personpath22
#153
ramafahad
opened
4 days ago
0
fixed bugs of deprecated numpy types and .DS_STORE issue
#152
yehlin-df
closed
3 weeks ago
0
Error: Tracker predicts the same ID more than once in a single timestep
#151
WJ-Huang
opened
2 months ago
0
Feature/per 105 customize track eval repo to fit evaluation of our trackers
#150
Hong-Pan
closed
4 months ago
0
Error while running the python3 scripts/run_rob_mots.py --ROBMOTS_SPLIT train --TRACKERS_TO_EVAL STP --USE_PARALLEL True --NUM_PARALLEL_CORES 4
#149
shivakarnati
closed
5 months ago
1
Adding a support to CholecTrack20 dataset on a new branch cholec
#148
nwoyecid
opened
5 months ago
0
Negative metrics value
#147
almex-m-bulbul
closed
5 months ago
1
Too Low HOTA and IDF1
#146
H1-KEY
closed
5 months ago
1
Numpy version
#145
setarekhosravi
closed
9 months ago
1
Increment the version number
#144
Scott-Percipient
closed
9 months ago
0
Is there any difference in the IDSW metric between TrackEval and motmetrics?
#143
mesllo-bc
closed
5 months ago
1
Evaluation of tracks where not a single object was tracked
#142
jokober
closed
5 months ago
7
I encountered an error and don't know how to solve it
#141
zeng121
closed
5 months ago
1
Upgrade modules to versions that will install on Ubuntu and expose exception message
#140
Scott-Percipient
closed
11 months ago
0
3D MOT code
#139
chumpblocckami
opened
1 year ago
2
how to calculate HOTA for DeepSORT?
#138
tianyufang1958
closed
5 months ago
1
The difference between MOT and MOTS evaluation on HOTA metrcis
#137
NanH5837
closed
5 months ago
1
How are 2d bounding boxes labeled in kitti ?
#136
kshashankrao
closed
5 months ago
1
Evaluation of class "cyclist"
#135
srinuSDV
opened
1 year ago
0
About the confused relationship between DetA and predict IDs.
#134
HELLORPG
opened
1 year ago
2
seqinfo.ini error
#133
BoscoShin
opened
1 year ago
1
Tracker evaluate on KITTI dataset but has format error
#132
chenx1e
opened
1 year ago
1
regarding track id matching and tracker results
#131
benhgm
opened
1 year ago
0
How can I further understand the meaning of "Metric arguments"?
#130
xieyunjiao
opened
1 year ago
0
a method to modify the result file to csv
#129
unfollow7
opened
1 year ago
0
gt_fol path fix
#128
noyzzz
opened
1 year ago
0
Implementation of ai city challenge?
#127
johnnynunez
opened
1 year ago
0
How to evalute test set of kitti 2D
#126
gehadza
closed
5 months ago
2
How to evaluate test data without ground truth?
#125
gehadza
closed
5 months ago
1
Align with official BDD100K submission format
#124
nicolamarinello
opened
1 year ago
0
How to directly specify sequences to eval and their number of timesteps using `SEQ_INFO`
#123
mikel-brostrom
opened
1 year ago
2
How is ground truth derived?
#122
adelinawong
closed
5 months ago
3
Error when running the example for MOTChallenge
#121
siiigrid
opened
1 year ago
1
Support BEV or 3D tracking?
#120
PhyllisH
opened
1 year ago
1
What is the purpose of <conf> column?
#119
slavastar
closed
5 months ago
2
Is there a plan to visualize tracking results?
#118
yuanyue668
opened
1 year ago
0
Replace deprecated NumPy aliases with builtin types
#117
piotlinski
opened
1 year ago
8
Fix --OUTPUT_FOLDER as List instead of String
#116
vexagonverp
opened
1 year ago
0
--OUTPUT_FOLDER param input in as List instead of String
#115
vexagonverp
opened
1 year ago
1
AttributeError: module 'numpy' has no attribute 'float'
#114
mikel-brostrom
opened
1 year ago
1
Update numpy to 1.20.1
#113
AtomScott
opened
1 year ago
0
Fix typo in readme.md
#112
AtomScott
opened
1 year ago
0
What is difference between HOTA and HOTA(0) ?
#111
ssbilakeri
closed
5 months ago
2
Is there an easy way of evaluating on half of MOT17?
#110
mikel-brostrom
opened
1 year ago
1
What is the evaluation format of OWTA?
#109
HangZheng66
opened
1 year ago
0
evaluated results of ByteTrack and Track eval
#108
ssbilakeri
opened
1 year ago
0
🐛 HOTA: fix numpy error when either ground truth or tracks are empty.
#107
OlivierLDff
opened
1 year ago
0
Different HOTA for the same ground-truth (DPM / FRCNN / SDP) of MOT17
#106
sieumap43
closed
1 year ago
2
'SEQMAP_FILE': None; if not None what exactly should be assigned
#105
ssbilakeri
opened
1 year ago
0
Should MOTP be higher or lower?
#104
ignaciomendizabal
opened
1 year ago
2
Next