A matlab implementation of GOTURN. Rather Fast in CPU(20FPS).
GOTURN appeared in this paper:
Learning to Track at 100 FPS with Deep Regression Networks, European Conference on Computer Vision (ECCV), 2016
GOTURN addresses the problem of single target tracking: given a bounding box label of an object in the first frame of the video, we track that object through the rest of the video.
Here is a brief overview of how GOTURN works:
Download from google drive or BaiduYun
or
Convert caffe models to matconvnet by yourself.
- Download caffemodel
cd <GOTURN_matconvnet>
cd model
wget http://cs.stanford.edu/people/davheld/public/GOTURN/trained_model/tracker.caffemodel
- MATLAB
cd model
Converter_caffe_matconvnet
make_goturn_net
- Compile matconvnet
cd <GOTURN_matconvnet>
git clone https://github.com/vlfeat/matconvnet.git
cd matconvnet
run matlab/vl_compilenn ;
- Download VOT2015
cd <GOTURN_matconvnet>
cd data
wget http://data.votchallenge.net/vot2015/vot2015.zip
unzip -n vot2015.zip -d ./VOT15
- Track in MATLAB
cd <GOTURN_matconvnet>
cd track
show_tracker_test();
**Ubuntu16.04 MATLAB2016b matconvnet-1.0-beta24 **
CPU: Intel(R) Xeon(R) CPU E5-2630 v4 @ 2.20GHz
GPU: GTX 1080ti
index | Name | CPU(fps) | GPU(fps) |
---|---|---|---|
Video: 1 | bag | 19.752 | 87.459 |
Video: 2 | ball1 | 18.36 | 92.336 |
Video: 3 | ball2 | 18.841 | 100.206 |
Video: 4 | basketball | 20.557 | 89.386 |
Video: 5 | birds1 | 18.177 | 96.058 |
Video: 6 | birds2 | 20.574 | 95.602 |
Video: 7 | blanket | 20.751 | 98.822 |
Video: 8 | bmx | 17.976 | 79.232 |
Video: 9 | bolt1 | 19.737 | 99.359 |
Video: 10 | bolt2 | 19.728 | 99.905 |
Video: 11 | book | 19.929 | 96.446 |
Video: 12 | butterfly | 19.1 | 94.681 |
Video: 13 | car1 | 18.178 | 67.749 |
Video: 14 | car2 | 20.272 | 96.047 |
Video: 15 | crossing | 19.109 | 92.823 |
Video: 16 | dinosaur | 19.865 | 92.37 |
Video: 17 | fernando | 19.033 | 78.656 |
Video: 18 | fish1 | 20.49 | 95.689 |
Video: 19 | fish2 | 20.609 | 97.977 |
Video: 20 | fish3 | 20.411 | 94.673 |
Video: 21 | fish4 | 19.897 | 96.342 |
Video: 22 | girl | 20.181 | 94.281 |
Video: 23 | glove | 19.159 | 94.516 |
Video: 24 | godfather | 19.934 | 101.74 |
Video: 25 | graduate | 20.496 | 99.465 |
Video: 26 | gymnastics1 | 20.224 | 95.631 |
Video: 27 | gymnastics2 | 17.944 | 94.272 |
Video: 28 | gymnastics3 | 17.924 | 88.965 |
Video: 29 | gymnastics4 | 19.763 | 93.196 |
Video: 30 | hand | 20.212 | 98.163 |
Video: 31 | handball1 | 20.486 | 99.822 |
Video: 32 | handball2 | 20.04 | 94.673 |
Video: 33 | helicopter | 19.653 | 87.207 |
Video: 34 | iceskater1 | 20.101 | 96.558 |
Video: 35 | iceskater2 | 19.575 | 91.674 |
Video: 36 | leaves | 18.942 | 99.633 |
Video: 37 | marching | 18.284 | 88.932 |
Video: 38 | matrix | 19.811 | 95.412 |
Video: 39 | motocross1 | 20.769 | 96.067 |
Video: 40 | motocross2 | 16.784 | 78.495 |
Video: 41 | nature | 19.877 | 88.762 |
Video: 42 | octopus | 16.946 | 75.872 |
Video: 43 | pedestrian1 | 21.029 | 98.191 |
Video: 44 | pedestrian2 | 20.381 | 96.385 |
Video: 45 | rabbit | 20.271 | 97.809 |
Video: 46 | racing | 19.932 | 96.254 |
Video: 47 | road | 18.91 | 96.028 |
Video: 48 | shaking | 20.484 | 95.895 |
Video: 49 | sheep | 21.3 | 101.453 |
Video: 50 | singer1 | 20.772 | 95.205 |
Video: 51 | singer2 | 20.427 | 91.472 |
Video: 52 | singer3 | 20.131 | 83.639 |
Video: 53 | soccer1 | 20.029 | 89.014 |
Video: 54 | soccer2 | 19.094 | 96.23 |
Video: 55 | soldier | 17.937 | 92.634 |
Video: 56 | sphere | 20.265 | 92.279 |
Video: 57 | tiger | 19.663 | 96.815 |
Video: 58 | traffic | 18.328 | 97.919 |
Video: 59 | tunnel | 20.135 | 98.278 |
Video: 60 | wiper | 20.404 | 97.771 |
If you have problem, email @foolwood.