Project idea · ungraded

Photo Filters with NumPy

Build a small photo-filter program using only NumPy array operations: greyscale, brightness and contrast, blur, edge detection and a vignette — applied to your own photos.

You will use

Milestones

Build the smallest working version first. Ticks are saved in this browser only.

01Load a photo into a NumPy array (with Pillow) and print its shape and dtype.
02Write greyscale, brightness and contrast filters with vectorised arithmetic, clipping to 0–255.
03Write a box blur by averaging shifted copies of the image.
04Write a simple edge filter from differences between neighbouring pixels.
05Add a vignette with a broadcast distance-from-centre mask, and save the results side by side.

Stretch goals

Write a sepia filter as a 3×3 colour matrix applied with matrix multiplication.
Compare your blur's speed with a Python double loop.
Build a command-line tool that chains filters: `python filters.py in.jpg out.jpg grey blur vignette`.

Data

Your own photos. Pillow only loads and saves the files; the filters themselves use NumPy (broadcasting guide).

Run it yourself

pip install numpy pillow
python filters.py photo.jpg