Detect-Track
Detect-Track copied to clipboard
Code release for "Detect to Track and Track to Detect", ICCV 2017
Luo, Hao, Wenxuan Xie, Xinggang Wang, andWenjun Zeng. "Detect or Track: Towards Cost-Effective Video Object Detection/Tracking". AAAI, 2019, 8. Is this also your work. I will be appreciate if you...
I can't open the link to download the trained model .I am in China ,angbody has the same problem?
@feichtenhofer Hi,I want to know how can I train a end2end model? Look forward to your reply.
Hello~ How can I get teh meta_vid.mat which used by imdb_from_ilsvrc15vid.m?
I have run the test code successfully, but I still have some doubts  1.To get the "Detect" results by "ImageNet CLS models" or Detect models"? 2.What is the difference...
Hi, I was trying to test this model on ImageNet VID (no modification to the code) and I used the trained models linked in the repo homepage (like [this one](http://ftp.tugraz.at/pub/feichtenhofer/detect-track/models/ResNeXt101-DT-ilsvrc-vid.caffemodel)...
Hello, how can I test this on my own video to see the performance? Thanks
when deploy the model, no groundtruth, so no trackids, so how to compute [rois_disp](https://github.com/feichtenhofer/Detect-Track/blob/master/models/rfcn_prototxts/ResNeXt-101L_ILSVRCvid_corr/test_track_reg.prototxt#L16) in the test network, use all of the rois?
I need help. I don't know how to get the input parameters of rfcn_test() and so on. What should I do set the parameters?
Hi,guys,I'm new here. I've downloaded this code, but do not know how to run this code for test as the author said. Could anyone explain explicitly about the pip line...