CMT (Consensus-based Matching and Tracking of Keypoints for Object Tracking) is a novel keypoint-based method for long-term model-free object tracking in a combined matching-and-tracking framework. Details can be found on the project page and in our publication. The Python implementation in this repository is platform-independent and runs on Linux, Windows and OS X.

CMT is freely available under the 3-clause BSD license, meaning that you can basically do with the code whatever you want. If you use our algorithm in scientific work, please cite our publication

@inproceedings{Nebehay2015CVPR,
    author = {Nebehay, Georg and Pflugfelder, Roman},
    booktitle = {Computer Vision and Pattern Recognition},
    month = jun,
    publisher = {IEEE},
    title = {Clustering of {Static-Adaptive} Correspondences for Deformable Object Tracking},
    year = {2015}
}

• Python
• OpenCV-Python (>= 2.4, < 3)
• NumPy
• SciPy
• optional: ipdb (for debugging the code)

Note for Windows users: if you are unable to read video files, please follow this suggestion: http://stackoverflow.com/questions/11699298/opencv-2-4-videocapture-not-working-on-windows

usage: run.py [-h] [--challenge] [--preview] [--no-preview] [--no-scale]
               [--no-rotation] [--bbox BBOX] [--pause] [--output-dir OUTPUT]
               [--quiet]
               [inputpath]

• inputpath The input path.
• -h, --help show help message and exit
• --challenge Enter challenge mode.
• --preview Force preview
• --no-preview Disable preview
• --no-scale Disable scale estimation
• --with-rotation Enable rotation estimation
• --bbox BBOX Specify initial bounding box. Format: x,y,w,h
• --pause Pause after each frame
• --skip N Skips N frames of the video input
• --output-dir OUTPUT Specify a directory for output data.
• --quiet Do not show graphical output (Useful in combination with --output-dir).

Press any key to stop the preview stream. Left click to select the top left bounding box corner and left click again to select the bottom right corner.

When using a webcam, no arguments are necessary:

python run.py

When using a video, the path to the file has to be given as an input parameter:

python run.py /home/cmt/test.avi

It is also possible to specify the initial bounding box on the command line.

python run.py --bbox=123,85,60,140 /home/cmt/test.avi

Use a sequence of numbered image files as an input:

python run.py sequence_dir/{:08d}.jpg

Here, {:08d} is a python format string that is expanded to 00000001.jpg, 00000002.jpg, etc.