mhttx2016/multi-camera-object-tracking-via-transferring-representation-to-top-view

Multi-Camera Object Tracking via Deep Metric Learning

multi-camera-object-tracking-via-transferring-representation-to-top-view's Issues

How to train on my own dataset?

Hi.
Nice work!
Could you please tell me how to retrain the network and how to train on other dataset?
Thanks a lot!

Could not run the result video files.

I am running the code on a CPU and after placing all the required files in the directories. The code is running fine but I can not play the result videos it gives an error of "Could not demultiplex stream". I am giving the correct frame size which is (360,288) for 4-people video sequences.
Can anyone help?

training code

Hello， Thank you for your contribution. could you provide training code?

train my own datasets

thanks for your code

how to train my own datasets?can you tell me ?

Process takes hours for few seconds videos

No doubt the solution is working fine.
I am using the same example of 4 videos of 2 minutes 36 seconds each. But it is taking around 40 minutes to complete the whole process and even though i am using GPU on google collab.
Can you please help on solving this issue?

dataset for object detection

Hey @Mhttx2016 thanks for your work.
I wanted to know which dataset have you trained your object detector and tracker on?
If you have used the EPFL 4 person dataset for training the tracker model, may I ask if you have used the camera calibration?

Permission to download data

Can anyone give me an account which has access to download data or anyone can grant me permission to download data ?

Real time detection

hello i want to use multiple webcams as input how can i do that i tried adding ip adrresses in the cap list but it didnt work please help thank you

create file .pickle

Hi ..
How can I create file .pickle for other video demo?
Please help me

Thanks!

It doesn't work when test on my own datasets

Does anyone complete the project and test on your own datasets?
Why are only two ID tracked when test on my own datasets? but there are seven persons in my datasets.
The ID switch is also very terrible，so , can someone tell me how to change the project?

Do you submit any paper by this method?

Do you submit any paper by this method?
I want to know the precise algorithm

Can any one share the Requirements file

can any one please share me the requirements file on which this code was built successfully.

where is your dataset_3000_0106.pickle file? thanks

runing error

hi ...
when i run the code i got this following error :

currentUrl
currentUrl /home/mohi/Multi-Camera-Object-Tracking-via-Transferring-Representation-to-Top-View-master/box2vec
parentUrl /home/mohi/Multi-Camera-Object-Tracking-via-Transferring-Representation-to-Top-View-master
video_files:
['data/train/lab/4p-c0.avi', 'data/train/lab/4p-c1.avi', 'data/train/lab/4p-c2.avi', 'data/train/lab/4p-c3.avi']
detection_files:
['data/train/lab/4p-c0.pickle', 'data/train/lab/4p-c1.pickle', 'data/train/lab/4p-c2.pickle', 'data/train/lab/4p-c3.pickle']
Traceback (most recent call last):
File "master.py", line 284, in
demo(save=True)
File "master.py", line 188, in demo
caps, detections, video_length = get_caps_and_pickles(video_files, detection_files)
File "master.py", line 26, in get_caps_and_pickles
detections.append(np.load(detection_file))
File "/home/mohi/.local/lib/python3.6/site-packages/numpy/lib/npyio.py", line 384, in load
fid = open(file, "rb")
FileNotFoundError: [Errno 2] No such file or directory: 'data/train/lab/4p-c0.pickle'

help to solve this issue thanks.

Different areas

Great project, the tracking, and reid work well when cameras are looking at the same area. Would like to ask whether anyone has figured out a way to make it work when cameras track different spaces?

where is dataset and train model file ?

Whenever click on train data it give error 👍

and whenever click on trained model : same error 👍

Whenever click on [detection file for demo] : Same error :