View on GitHub

dlvot

Download VOT Challenge sequences.

dlvot

Installation

  1. Download the latest version of dlvot
  2. Install the requests and wget packages for Python:
    $ pip install requests wget
    

Usage

At its most basic, just tell dlvot which data sets to download. This example will download the 2017 and 2018 VOT data sets:

$ dlvot 2017 2018

On Windows:

C:\Users\you> py dlvot 2017 2018

Run dlvot --help for complete information.

Note for Windows Users

On Windows, you must run dlvot with elevated privileges. Otherwise, you will get an OSError exception when the application tries to create a symlink.

How dlvot Works

VOT reuses sequences, just updating the tags and ground truth data. dlvot takes advantage of this to save space on your system. Only one copy of the frame data is kept. Symlinks mimic the directory structure typically expected by the VOT toolkit.

The root download directory is ~/Videos/vot/. Within this directory is a sequences/color/ subdirectory. At this time, color is the only supported channel. Also in the root directory is a subdirectory for each data set year. So, the directory structure looks like this:

~/Videos/vot/
  sequences/color
    bicycle/
      00000001.jpg
      00000002.jpg
      <rest of the images>
  2013/
    bicycle/
      color -> ~/Videos/vot/sequences/color/bicycle
      groundtruth.txt
      <tag files>
  2014
    bicycle/
      color -> ~/Videos/vot/sequences/color/bicycle
      groundtruth.txt
      <tag files>

dlvot creates any necessary directories, downloads the archives from VOT, unpacks the archives, creates the symlinks, and deletes the archives.