movievis

movievis is a python cli tool, to create still visualizations of movies (displaying color/time evolution) dimension and multi-color abstractions from images.

I tried to optimize for performance a little using multiprocessing pools, but there's plenty of room for optimization. The k-means is pretty hungry, the rest is up to the image generation done by ffmpeg.

Usage

Receive the options by calling mv.py --help

movievis [-t TYPE] [--force] imagefile|imagedir
movievis [-t TYPE] [--framecount=400] [--keep] moviefile
blockwidth, blockheight, and tilewidth are additional options for the according type of visualization

Options:
  -h, --help            show this help message and exit
  -t TYPE, --type=TYPE  Type of the visualization - blocks (default), pie,
                        tiles (only moviefiles)
  --blockheight=BLOCKHEIGHT
                        Block height (default: 150)
  --blockwidth=BLOCKWIDTH
                        Block width (default: 1)
  --tilewidth=TILEWIDTH
                        Tile width (default: 1)
  -c FRAMECOUNT, --framecount=FRAMECOUNT
                        Number of frames extracted from moviefile (default:
                        400)
  -f, --force           Force new calculation of colors for target (obsolete
                        for type=tiles)
  -k, --keep            Keep generated thumbs

Credits and dependencies

movievis uses

• ffmpeg to grab frames from video files, supported codecs depend on ffmpeg of course
• some python libs: matplotlib, PIL, multiprocessing, cPickle (make sure these are installed)
• k-means clustering to identify key colors. The k-means implementation for images is taken from Ze'ev Gilovitz blog post and on github. (included)
• an ffmpeg wrapper by Senko Rasic et.al. found here (included)