# Project 4: Toni Stachewicz, Niklas Hoffmann - WritePad Handwriting Recognition

[presentation](presentation.pdf)

![](screenshot.jpg){width=400, style="float:right;margin:20px"}

## Abstract

In an era of machine learning, recognizing handwritten text becomes a solveable task. When looking on GitHub, one can find various kinds of implementations of handwritten text or character recognition. Their objective is to provide a native app running on the operating system or detect single symbols with trained neural nets. However, finding an implementation which is open source on the one hand and has a good recognition rate on the other hand. [WritePad](http://www.phatware.com/index.php?q=page/writepad/ipad/de#) by PhatWare fullfills these requirements. Our application integrates the WritePad handwriting library into the lively ecosystem.


The repository of the web server and WritePad integration can be found [on github](https://github.com/NHoff95/WebDev17_18_HandwritingRecognition).


## Screencast

<video autoplay controls width=800><source src="screencast.mp4" type="video/mp4"></video>


## Project

### Failed Attempts

- [Grail Handwriting recognition](https://lively-kernel.org/lively4/lively4-core/demos/grail-handwriting.html)
- [Tesseract as JS Library (OCR)](https://lively-kernel.org/lively4/lively4-core/demos/tesseract_hand.html)
- MyScript (Paid Service: 2000 Requests per month free)
- [ocrad.js - ocr using emscripten](https://antimatter15.com/project/ocrad-js/)
- Windows App (Awesome Recognition, not accessible from outside)

### Successes

- [MorphicJS (Windows App)](https://github.com/timfel/morphicjs-pen)
- [Hyper Lively (Astrid Thomschkes Master Thesis)](http://athomschke.github.io/hyperlively/src/client/index.html)

### Task: Go wide!

- Explore many possibilities (for now)
- No polishment
- Target immediate handwriting recognition