Create a mobile handwritten Hangul translation app  

Create a mobile application leveraging TensorFlow that will recognize and translate handwritten Korean characters

Last updated | By Paul Van Eck


The Korean written language has thousands of unique characters that make up its words. In this pattern, you’ll learn how to use TensorFlow and the Watson Language Translator to make an Android app capable of recognizing and translating the Korean words you draw on your devices.


Hangul, the Korean alphabet, has 19 consonants and 21 vowel letters. Combinations of these letters give a total of 11,172 possible Hangul syllables and characters. However, only a small subset of these are typically used.

In this pattern, you’ll go through the process of generating your own Korean training data and then train a TensorFlow model to classify some of the most common Hangul handwritten characters. You’ll then build and run an Android application, where you are able to draw Korean characters on your mobile devices and have the characters recognized by using the trained model. Next, you’ll form Korean words or sentences in the application, which you can then translate by using the Watson Language Translator service.


  1. The user downloads several Korean fonts to use for data generation.
  2. The images generated from the fonts are fed into a TensorFlow model for training.
  3. The user draws a Korean character on their Android device.
  4. The drawn character is recognized by using the previously trained TensorFlow model and the Android TensorFlow Inference interface.
  5. A string of the classified Korean characters is sent to the Watson Language Translator service to retrieve an English translation.

Related Blogs

Two “edgy” AI TensorFlow models for you!

The global Call for Code is well underway, we want to share some visual recognition models which could help you. These AI models can operate on the edge, which could be particularly useful for this years’ theme: disaster preparedness. How could visual recognition help in relief work? From satellite and drone imagery analysis, to classifying...

Continue reading Two “edgy” AI TensorFlow models for you!

Leveraging the power of AI at Unite Berlin

Last week, from June 19 – 21, we were at Unity’s premiere in Berlin: Unite 2018. This conference brought together Unity’s video game and development community. Unity touches 770 million gamers all over the world and is the market leader for consumer AR and VR use cases and is also rapidly emerging as the market...

Continue reading Leveraging the power of AI at Unite Berlin

Related Links

Deep MNIST for Experts

Tutorial for creating and training a convolutional neural network to recognize handwritten digits.