Back to Projects
Machine Learning

Handwriting Recognition System

FlutterMLOCRCanvas APIGesture Detection
Aa

About this Project

Designed and implemented a sophisticated handwriting recognition system for NOTEorious Notes, featuring advanced OCR technology, machine learning algorithms, and real-time text processing capabilities for seamless note digitization.

Highlights
01

Write like on paper

A custom canvas captures every stroke with its pressure and timing, so writing feels natural on glass.

02

Instant conversion

Strokes are processed through machine learning models that turn handwriting into editable text as you write.

03

Seamless mode switching

One tap toggles between typing and handwriting, and both feed into the same note.

04

Tuned for real hands

Gesture detection is tuned on real, messy handwriting rather than clean lab samples.

The challenge

Real handwriting is chaotic. Letters connect, sizes drift, and every hand is different. The system needed careful tuning of stroke segmentation and model thresholds before it stopped guessing wrong on everyday notes.

The outcome

The final system ships inside NOTEorious and lets note-takers move between pen and keyboard without thinking about it.

Project Details

Category

Machine Learning

Completed

February 10, 2025