An inclusive, real-time handwriting-to-text system — no keyboard required.
Typing isn’t designed for everyone. People with motor disabilities or limited hand mobility often struggle with keyboards. This project introduces a natural, accessible way to communicate: through handwriting input instead of traditional typing.
Keyboardless Handwriting Input allows users to “write” in the air, on a touchscreen, or with a stylus — and see their handwriting instantly translated into text.
- ✍️ Converts natural handwriting gestures into typed text
- 🎯 Tracks finger or stylus movements using computer vision
- ⚡ Recognizes and translates handwriting into text in real time
Unlike standard typing tools, this project is:
- Accessible: Designed for users with motor impairments
- Versatile: Works on phones, tablets, smartwatches, and even in AR/VR
- Natural: Mimics real handwriting for faster, more comfortable text input
- No physical keyboard needed
- Write on touchscreens or “in the air”
- Real-time handwriting recognition powered by machine learning
- Works across diverse scripts and input styles
Who Benefits:
- Individuals with accessibility needs
- Students and professionals in hands-busy environments
- Creatives and multilingual users who prefer pen-like input
Use Cases:
- Note-taking, form-filling, creative expression, and accessible communication
- Input in multiple languages or non-Latin scripts with ease
- MediaPipe Hands: Real-time hand landmark detection (via webcam)
- HTML5 Video: Webcam streaming in the browser
- JavaScript + TensorFlow.js: Extracts finger positions client-side
- OpenCV (Python): Image preprocessing (grayscale, resize)
- TensorFlow / PyTorch: Character recognition using MNIST/EMNIST
- Flask (Python): Handles API calls and model inference
- REST API: Connects frontend input to backend prediction logic
- React.js or Vanilla JS
- Canvas API: Visualizes the user's handwriting path in real time
git clone https://github.com/anunjinb/Keyboardless-Handwriting-Input.git
cd Keyboardless-Handwriting-Input