+------------------------------------------------------------------------------+
An image compression utility that leverages quadtree data structures to reduce image file sizes while maintaining visual quality through recursive spatial partitioning.
How It Works
The algorithm progressively subdivides an image into four quadrants at each level. Subdivisions continue if the depth limit hasn’t been exceeded or the quadrant contains excessive detail.
+------------------+ +--------+--------+
| | | | |
| | | Q1 | Q2 |
| Original | ---> +--------+--------+
| Image | | | |
| | | Q3 | Q4 |
+------------------+ +--------+--------+
Further subdivision based on detail:
+--------+--------+ +----+---+--------+
| | | | a | b | |
| Q1 | Q2 | ---> +---+---+ Q2 |
+--------+--------+ | c | d | |
| | | +---+---+--------+
| Q3 | Q4 | | | |
+--------+--------+ | Q3 | Q4 |
+--------+--------+Algorithm Flow
+-------------------+
| Load Image |
+--------+----------+
|
v
+-------------------+
| Calculate Detail |
| for Region |
+--------+----------+
|
+--------------+--------------+
| |
v v
+-------------------+ +-------------------+
| Detail > Threshold| | Detail <= Threshold|
| AND Depth < Max | | OR Depth >= Max |
+--------+----------+ +--------+----------+
| |
v v
+-------------------+ +-------------------+
| Subdivide into | | Store average |
| 4 quadrants | | color for region |
+--------+----------+ +-------------------+
|
v
+-------------------+
| Recurse on each |
| quadrant |
+-------------------+Quadtree Data Structure
[Root]
/ | \ \
/ | \ \
[Q1] [Q2] [Q3] [Q4]
/|\
/ | \
[a][b][c][d]
Each node stores:
- Bounding box coordinates
- Average color (RGB)
- Detail level
- Child nodes (if subdivided)Tech Stack
- Language: Python
- Image Processing: PIL (Python Imaging Library)
Features
- Recursive quadtree-based compression algorithm
- Visual representation of compression progress (animated GIFs)
- Adjustable depth level and detail threshold parameters
- Generates both compressed output and visualization with quadrant boundaries
Built with @tanneryork