Project library
Handwriting Legibility Scoring
Explore how confidently a model recognizes your handwritten digits.
STAR Summary
- Situation
- My wife says my handwriting is hard to read. I wanted an objective score.
- Task
- Build a handwritten-digit recognition model and use its confidence to explore legibility scoring.
- Action
-
- Built three digit-recognition models, moving from a simple baseline to a CNN.
- Trained on MNIST (60,000 digits) and selected the best model.
- Deployed it behind a serverless scoring API for the live demo.
- Result
-
- The best model reached ~99% test accuracy on MNIST (see notebook).
- On a small personal handwriting set, accuracy was ~75.6% (not a benchmark; see notebook).
- My wife was right.
Handwriting Rating
Draw a digit to see how the model recognizes it.
Open full demo
Draw a digit (0-9) or load a sample, then compare the model's guesses.
- Draw on the canvas, or use the sample buttons to try Daniel’s handwriting.
- Click “Rate digit” to submit.
- Compare all ten digit scores, ranked from highest to lowest.
- Select Clear, or press C, to draw another digit.
