PyraDox :page_with_curl:

PyraDox is a simple tool which helps in document digitization by extracting text information and masking of personal information with the help of Tesseract-ocr.

Currently Supports :-

• Aadhaar Card is a 12-digit unique identity number that can be obtained voluntarily by residents or passport holders of India, based on their biometric and demographic data. The data is collected by the Unique Identification Authority of India (UIDAI), a statutory authority established in January 2009 by the government of India.

Installation

Tesseract-ocr

This tools need tesseract-ocr engine. Help yourself with this --

• https://github.com/tesseract-ocr/tesseract/wiki

Windows

Install tesseract using windows installer available at :

• https://github.com/UB-Mannheim/tesseract/wiki

Linux

Tesseract is available directly from many Linux distributions. The package is generally called 'tesseract' or 'tesseract-ocr' - search your distribution's repositories to find it. Thus you can install Tesseract 4.x and it's developer tools on Ubuntu 18.x bionic by simply running:

sudo apt install tesseract-ocr
sudo apt install libtesseract-dev

Refer here for more on installation on all other systems.

macOS

Homebrew

To install Tesseract run this command:

brew install tesseract

Dependency

Use the package manager pip to install requirements.

pip install -r requirements.txt

Having hard time with pyt

Add path if pytesseract is unable to find Tesseract-ocr path. stackoverflow

pytesseract.pytesseract.tesseract_cmd = r'C:\Users\USER\AppData\Local\Tesseract-OCR\tesseract.exe'

Usage

Initialisation & Configuration

psm modes

from Aadhaar import Aadhaar_Card
config = {'orient' : True,   #corrects orientation of image default -> True
          'skew' : True,     #corrects skewness of image default -> True
          'crop': True,      #crops document out of image default -> True
          'contrast' : True, #Bnw for Better OCR default -> True
          'psm': [3,4,6],    #Google Tesseract psm modes default -> 3,4,6 
          'mask_color': (0, 165, 255),  #Masking color BGR Format
          'brut_psm': [6]    #Keep only one for brut mask (6) is good to start
          }

obj = Aadhaar_Card(config)

A. Validate Aadhaar card numbers using Verhoeff Algorithm.

obj.validate("397788000234") #Binary Output 1|0

B. Extract Aadhaar Number from image

aadhaar_list = obj.extract("path of input image") #supported types (png, jpeg, jpg)

C. Mask Aadhaar number card for given Aadhaar card number #Binary Output 1|0

flag = obj.mask_image("path of input image", "path of output image", aadhaar_list) #supported types (png, jpeg, jpg)

D. Brut Mask any Readable Number from Aadhaar (works well on low res, bad quality images)

obj.mask_nums("path of input image", "path of output image") #supported types (png, jpeg, jpg)

PyraDox-API

Built with flask

Find Usefull Examples of Request - Response api_samples

defaults_url = http://localhost:9001
headers = {'content-type': 'application/json'}

python app.py

A. Validate Aadhaar card numbers using Verhoeff Algorithm. url = '/api/validate'

request_json = {"test_number": 397788000234} 
response_json = {'validity': 0 } #0|1 -> invalid|valid

B. Extract Aadhaar Number from image. url = '/api/ocr'

request_json = {"doc_b64": base64_encoded_string}
response_json = {'aadhaar_list':['397788000234']} #enpty list if unable to find

C. Mask Aadhaar number card for given Aadhaar card number. url = '/api/mask'

request_json = {"doc_b64": base64_encoded_string, 'aadhaar': ['397788000234']}
response_json = {'doc_b64_masked':base64_encoded_string, 'is_masked': True} #if is_masked False then doc_b64_masked is None

D. Brut Mask any Readable Number from Aadhaar (works well on low res, bad quality images). url = '/api/brut_mask'

request_json = {"doc_b64": base64_encoded_string}
response_json = {'doc_b64_brut_masked': base64_encoded_string, 'mask_status': 'Done'}

E. Bonus :100: Complete Sample Pipeline. url = '/api/sample_pipe'

Usecase : Take an aadhaar card, extract its aadhaar number while checking number's validty, mask first 8 digits. If aadhaar number is not readable then mask possible numbers (brut mode) .

request_json = {"doc_b64": base64_encoded_string, "brut" : True}
response_json = {'doc_b64_masked':base64_encoded_string, 'is_masked': True,'mode_executed' : "OCR-MASKING", 'aadhaar_list':"All Possible Aadhar Numbers of 12 digits", 'valid_aadhaar_list':['Valid Aadhar Numbers Only']}

Docker

Build Your Own Image

docker build -t pyradox .
docker run -p 9001:9001 pyradox

Samples

Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.

Tasks

• [x] Finish Dockerfile
• [x] Add Badges
• [ ] Add Class Preprocessing
• [ ] Sample Website
• [ ] Push Docker image to hub
• [ ] Add Regex to extract Name, DOB, Gender.

Please make sure to update tests as appropriate.

License

Apache License 2.0

Notes

Sample Aadhar Cards are just samples taken from google search and not original documents.

while working on this project, I came across some good repos on github :yum: which I am listing below.

Aadhar Number Validator and Generator Aadhaar-Card-OCR

If there is anything totally unclear, or not working, please feel free to file an issue. reach out at Email :innocent:

If this project was helpful for you please show some love :star: