jutanke / cabbage Goto Github PK

i got a importerror:
Traceback (most recent call last):
File "test.py", line 2, in
from cabbage.MultiplePeopleTracking import execute_multiple_people_tracking
File "/home/wangmingkun/code/cabbage/cabbage/MultiplePeopleTracking.py", line 5, in
from cabbage.features.GenerateFeatureVector import pairwise_features
File "/home/wangmingkun/code/cabbage/cabbage/features/GenerateFeatureVector.py", line 4, in
from cabbage.features.deepmatching import DeepMatching
File "/home/wangmingkun/code/cabbage/cabbage/features/deepmatching.py", line 6, in
from pppr import aabb
ImportError: No module named 'pppr'

How can I solve it?

• MOTP
• MOTA

For that we need to reproduce their detection "ground-truth", e.g. run over the gt and the detection data, find the bb's with IoU > 0.5 and use those as gt data

In the file MultiplePeople Tracking, line 54, the code use ' reid = StackNet64×64(root)', but i acn not find the 'root'. is there a bug?

Thanks for your nice implementation. I have a question for the final part.
The referenced paper said that they estimate a smoothed trajectory from the detections that belongs to the same cluster, by using the code from Continuous energy minimization for multitarget tracking. Did you use the same method for finding the trajectories which are depicted in the readme file's figure? Can you add the code for that part as well ?!

nothing works and it reminds me how old are we

workaround to upload images

I had tested on the MOT16, but got very poor performance.
Is anyone successful on reproducing the numbers on paper?
Maybe we can discuss on improving the project and help each other. @jutanke

I am facing issues to import cabbage. It still shows up in pip but I don't know why I cant import it. Can you please help solve it.

Hi, your work is interesting, but I can't find how to train this model on MOT dataset. can you tell about more details？Thx!

Thanks for putting this repo online!
Really like the paper.

I had a question regarding running the code.
Seems like I cannot start the subprocess necessary for deepmatching.
Running on macOS Mojave and python 3.5.
Would be very glad about some feedback!

Thanks a lot!
André

@jutanke

Sorry.
I can't understand which file lead to do execution of tracking by using the detection data and the nerwork python file

If anyone know how to use this repo, please teach me

Is there an example to use the output of the execute_multiple_people_tracking to generate the tracking output? How to assign the ids to the persons in the image?

Thanks for your LMP code, do you have the pytorch version code of LMP?

File "/home/gyc/GYC_Python_Work/Multiple-People-Tracking-by-Lifted-Multicut-and-Person-Re-identification.py", line 16, in
execute_multiple_people_tracking(video_folder, X, Dt, video_name, dmax, settings_file)
File "/home/gyc/GYC_Python_Work/cabbage/MultiplePeopleTracking.py", line 57, in execute_multiple_people_tracking
gg.build(Dt, X, W, batch_size=batch_size)
File "/home/gyc/GYC_Python_Work/cabbage/MultiplePeopleTracking.py", line 95, in build
lookup = AABBLookup(Dt, X)
File "/home/gyc/GYC_Python_Work/cabbage/features/combined.py", line 40, in init
im = get_element(X[frame-1], (x,y,w,h), (W, H), True)
File "/home/gyc/GYC_Python_Work/cabbage/features/ReId.py", line 23, in get_element
I = resize(I, shape, mode='constant').copy()
File "/home/gyc/anaconda3/lib/python3.6/site-packages/skimage/transform/_warps.py", line 135, in resize
preserve_range=preserve_range)
File "/home/gyc/anaconda3/lib/python3.6/site-packages/skimage/transform/_warps.py", line 819, in warp
_clip_warp_output(image, warped, order, mode, cval, clip)
File "/home/gyc/anaconda3/lib/python3.6/site-packages/skimage/transform/_warps.py", line 570, in _clip_warp_output
min_val = input_image.min()
File "/home/gyc/anaconda3/lib/python3.6/site-packages/numpy/core/_methods.py", line 29, in _amin
return umr_minimum(a, axis, None, out, keepdims)
ValueError: zero-size array to reduction operation minimum which has no identity

Traceback (most recent call last):
File "my_run.py", line 14, in
execute_multiple_people_tracking(video_loc, X, Dt, video_name, dmax, settings_loc)
File "/home/shaun/cabbage/cabbage/MultiplePeopleTracking.py", line 57, in execute_multiple_people_tracking
gg.build(Dt, X, W, batch_size=batch_size)
File "/home/shaun/cabbage/cabbage/MultiplePeopleTracking.py", line 95, in build
lookup = AABBLookup(Dt, X)
File "/home/shaun/cabbage/cabbage/features/combined.py", line 40, in init
im = get_element(X[frame-1], (x,y,w,h), (W, H), True)
File "/home/shaun/cabbage/cabbage/features/ReId.py", line 23, in get_element
I = resize(I, shape, mode='constant').copy()
File "/home/shaun/anaconda3/envs/tensorflow/lib/python3.6/site-packages/skimage/transform/_warps.py", line 135, in resize
preserve_range=preserve_range)
File "/home/shaun/anaconda3/envs/tensorflow/lib/python3.6/site-packages/skimage/transform/_warps.py", line 819, in warp
_clip_warp_output(image, warped, order, mode, cval, clip)
File "/home/shaun/anaconda3/envs/tensorflow/lib/python3.6/site-packages/skimage/transform/_warps.py", line 570, in _clip_warp_output
min_val = input_image.min()
File "/home/shaun/anaconda3/envs/tensorflow/lib/python3.6/site-packages/numpy/core/_methods.py", line 29, in _amin
return umr_minimum(a, axis, None, out, keepdims)
ValueError: zero-size array to reduction operation minimum which has no identity

From the code i cannot find body part fusing.
Is the fusing step in provided h5 model?

there is nothing about the model in the internet "http://188.138.127.15:81/models/". What should I do about it? thanks a lot if you tell me .