strangerzhang / pyeco Goto Github PK

Hi, I have several questions about your evaluation on OTB

1. OTBdeep's backbone is ResNet50 or VGG16 ?
2. Could you please send me a copy of your OTB toolkit pyhon version ?

what is your test speed？why is it too slow when i test (less than 1fps)?

hi
first, tx for the repo.

Now when executing the demo I have the following error (the video passed is a mp4 of 6mins so about 10k frames), can you help please?

~/pyECO$ python bin/demo_ECO_hc.py --video_dir ~/Desktop/videos/1000FPSSlowMotion.mp4 
/home/tets/.local/lib/python2.7/site-packages/h5py/__init__.py:34: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.
  from ._conv import register_converters as _register_converters
Traceback (most recent call last):
  File "bin/demo_ECO_hc.py", line 66, in <module>
    main(args.video_dir)
  File "bin/demo_ECO_hc.py", line 18, in main
    height, width = frames[0].shape[:2]
IndexError: list index out of range

你好，我最近在练习一些想关滤波代码的python复现，参考了这个repo，很感谢！发现了一个小小的问题。
features.py中283行应改成
RR = patch[:, :, 2].astype(np.int32)
GG = patch[:, :, 1].astype(np.int32)
BB = patch[:, :, 0].astype(np.int32)

Does this code work on Windows or Linux? When I tried "python setup.py build_ext --inplace", it says "make sure that python modules winreg, win32api or win32con are installed".

Dear Author,

Is there any way to integrate it with a deep learning based object detector and supply multiple objects to be tracked across the frame like KCF with an ID using this approach ?

Thank You !

Can i take a look your test code on OTB100? Tks!

Hello, i am new to visual tracking, and i have already done the demo running,
But i don't know how to running it with my own video? what should i do?
thanks for a lot!

Sorry ,I'm s a rookie in keras & Neural network.
Pardom,can I ask you a question?

I am trying my best to make your codes runnable the whole afternoon,but unfortunatly I failed，can you Provide a list of dependent kits?
Would you plz make the usage more detailed so that anyone interested in your work would ultimate and contribute it.

Thanks in advance : )

Thanks for your work. I have tested your code and it works very well. I wonder if there are any sample augmentation methods that have been used during filter training. If there are, please tell me where I can find them.
Thanks.

dear zhang,
i am trying to reimplement your code on my cpu and as per your instructions i need cupy environment but it recomended for CUDA device only. and when i am trying to run it without cupy following errors occurs
ImportError: cannot import name '_gradient' from 'eco.features'
i am using windows..
and what modifications i should make in comment line 6 of features..
can your please elaborate them..

In file eco/scale_filter.py, function of _extract_scale_sample
if xs.min() > im.shape[1] or ys.min() > im.shape[0], etc. , then the code will crash in cv2.copyMakeBorder

Before asking any further error question I would like to learn if this project is portable to run on Jetson Nano because all my errors seems like architecture problems (_gradient.o _gradient.o.d) and also from mxnet-cuda version unmatching leaks

The speed of ECO_HC is about 30fps on gtx1060,but,ECO_HC with C++ is 80fps on gtx1060. what's wrong ?

I have studyed ECO for two weeks and tried running the code you provided. Many problems were solved. But the "Segmentation fault" bothered me and I can't solve it.
If you are free, Can you help me?

Thank you for the code you contributed. I have been studying the code for a while, and there are many places in the program that don't quite understand. Pleading the author to provide pseudo code or execution flow chart of the algorithm ECO.Thank you very much.And,Looking forward to your reply.

When I run the commond "python bin/demo_ECO_hc.py --video_dir path/to/video", print ModuleNotFoundError: No module named 'eco' . Could you help me ?