Skip to content

2022-AI-FullStack-Internship-TeamB/NCTE

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

NCTE

Note Cursive To E-note

(for Korean Handwritten Notes)

Upload pictures of handwritten notes and get results of converted text.
With this Android/ios Application, users can manage handwritten notes online.

1. Usage Example

demo 1 demo 2

2. Architecture

System Architecture

system architecture

AI Architecture

ai architecure

3. Installation Process

$ git clone https://github.com/2022-AI-FullStack-Internship-TeamB/NCTE.git
$ docker-compose up --build

**docker failed, just install 'requirements.txt' and execute 'runserver' from 'backend' directory

4. Acknowledgement and References

This project is based on several open-source repositories.
We used EasyOCR for detecting and recognizing text. We fine-tuned EasyOCR using training script from here and korean_g2.pth.
Dataset we used for fine tuning is part of Korean OCR data from AI hub.
We used the KoBART model for content summarization. One of the examples provided by kobart, kobart-sumarization, was referenced. It is summarized using a pretrained model of kobart-sumarization.
KeyBERT was also used to extract keywords. Sentence Transformers, a package for SBERT, and KoNLPy, a morpheme analyzer, were used.

5. Copyrights / End User Licensce

This project is not intended for commercial use, please do not use it for commercial purposes.

6. Contact Information

Name 곽미래 곽서진 고은서 이정현 최정민
Role Backend/DevOps Frontend Frontend Backend/AI AI
Github @miraekwak @SJ-Kwak @godori1012 @hyunihs @JAIBC

Releases

No releases published

Packages

No packages published

Contributors 4

  •  
  •  
  •  
  •