Quart

Image Segmentation Animation using QuadTree concepts.

Split the Image into four quadrants.
Split the quadrant with the highest error into four quadrants.
Repeat N times.

Images

Borders

Compression

With or without borders, the quadtree images achieve great compression, especially when using png encoding. Looking at the Mountain Images above, the original is a 1.51 MB jpg file (7.89 MB when converted to png), while the Quadtree Image with borders is a 333 KB png and the one without borders is a 160 KB png.

Video

Usage

usage: quad.py [-h] [-q QUALITY] [-b] [-au] [-mw MINWIDTH] [-mh MINHEIGHT] input output iterations

Quadtree Image Segmentation.

positional arguments:
  input                 Image to segment.
  output                Output filename.
  iterations            Number of segmentation iterations.

optional arguments:
  -h, --help            show this help message and exit
  -q QUALITY, --quality QUALITY
                        Quality of the output video. (0-10), 0 worst, 10 best.
  -b, --border          Add borders to subimages.
  -a, --audio           Add audio from the input file to the output file.
  -mw MINWIDTH, --minwidth MINWIDTH
                        Minimum width of the smallest image quadrant.
  -mh MINHEIGHT, --minheight MINHEIGHT
                        Minimum height of the smallest image quadrant.

Dependencies

numpy
tqdm
imageio
imageio-ffmpeg

pip install numpy tqdm imageio imageio-ffmpeg