OCR-KTP

This project aims to create an API that can scan and convert important data (NIK, Name, Place and Date of Birth) from a KTP image into text using PyTesseract Optical Character Recognition (OCR). In addition there is also a deep learning (YOLO) based KTP detector that can automatically crop the KTP image to improve OCR's ability to read text on images (this feature is still very imperfect and requires further development). Thanks to the developers who have developed most of the contents of this system before.

Prerequisites

• Flask

pip install flask

• Numpy

pip install numpy

• OpenCV

pip install opencv-python

• Pandas

pip install pandas

• PIL

pip install pillow

• PyTesseract

sudo apt-get install tesseract-ocr
sudo apt-get install tesseract-ocr-ind
pip install pytesseract

• TextDistance

pip install textdistance

Running the Program

Create virtual environment:

python -m venv env

Run virtual environment (for Windows user):

env\Scripts\activate.bat

Run virtual environment (for Unix user):

source ./env/bin/activate

To run the program, use the command below:

export FLASK_APP=app
flask run

or alternatively using this command:

python app.py

Request Parameter

Parameter	Data Type	Mandatory	Notes
image	Image Files	M	Foto KTP

Response Parameter

Parameter	Description
nik	NIK dari hasil OCR
nama	Nama dari hasil OCR
tempat_lahir	Nama tempat lahir dari hasil OCR
tgl_lahir	Tanggal lahir dari hasil OCR (DD-MM-YYYY)
time_elapsed	Waktu yang pemrosesan yang dibutuhkan (detik)

Success Response Example

{
    "error": false,
    "message": "Proses OCR Berhasil",
    "result": {
        "nik": "1234567890123456",
        "nama": "DENNY SEPTIAN",
        "tempat_lahir": "JAKARTA",
        "tgl_lahir": "10-10-1999",
        "jenis_kelamin": "LAKI-LAKI",
        "agama": "ISLAM",
        "status_perkawinan": "BELUM KAWIN",
        "pekerjaan": "PELAJAR/MAHASISWA",
        "kewarganegaraan": "WNI",
        "alamat": {
            "name": "DUSUN 1 OGAN 5",
            "rt_rw": "001/002",
            "kel_desa": "SUNGAI ARE",
            "kecamatan": "ALANG-ALANG LEBAR",
            "kabupaten": "OGAN ILIR",
            "provinsi": "SUMATERA SELATAN"
        },
        "time_elapsed": "6.306"
    }
}

Notes for KTP Detection using CNN

1. Create new folder, data/cnn
2. Run the program

Acknowledgments

• https://github.com/enningxie/KTP-OCR
• https://github.com/jeffreyevan/OCR