Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

Appearance settings

Kuzushiji translation using Object Detection and Prediction Machine Learning Algorithms: Python, TensorFlow, Keras

Notifications You must be signed in to change notification settings

morutkin/MLProject

Open more actions menu

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

13 Commits
 
 
 
 
 
 

Repository files navigation

Machine Learning Final Project: Kuzushiji Translation

Kuzushiji

Background on the Kuzushiji Challenge

  • 800-1900 CE: Japan used a writing system called Kuzushiji

  • Cursive Japanese writing style comprising of Chinese characters and Japanese hiragana characters

  • 99.99% of Japanese speakers cannot read texts which are more than 150 years old

  • Three million books holding important information about Japanese culture and history remain inaccessible to the general public

  • Due to lack of manpower, the National Institute of Japanese Literature (NIJL) has turned to the help of computers to translate these texts into Modern Japanese characters

Goal

To use machine learning to automatically translate Kuzushiji to modern Japanese texts in order to unlock classical Japanese literature and Japanese history contained in these books.

Kuzushiji with the contemporary traslations on the side

Solution

I used a 2-step solution where I first use CenterNet neural networks to detect characters within a page, followed by Convoluted Neural Networks with a ResNet backbone to then classify each character to a modern Japanese character.

The kaggle competition can be found here: https://www.kaggle.com/c/kuzushiji-recognition/overview

About

Kuzushiji translation using Object Detection and Prediction Machine Learning Algorithms: Python, TensorFlow, Keras

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published
Morty Proxy This is a proxified and sanitized view of the page, visit original site.