thai-personal-card-extract 1.3

Library for extract infomation from thai personal identity card

ThaiPersonalCardExtract

Library for extract infomation from thai personal identity card. imprement from easyocr and tesseract

New Feature v1.3 🎁

• Increase performance.
• Support Thai Driving License (Beta) สามารถสกัดข้อมูลจากภาพถ่ายใบขับขี่ได้บางรูปแบบ เนื่องจาก กรมทางขนส่งทางบก มีรูปแบบบัตรหลากหลายรูปแบบ และแต่ละรูปแบบมีตำแหน่งข้อมูลที่แตกต่างกัน จึงทำให้ประสิทธิภาพต่ำ
• ปรับเปลี่ยนรูปแบบไฟล์ระบบ
• Support Thai Government Lottery (Further works)

Examples

Real image file.

wrapPerpective image crop.

keypoint of image detected.

Resutls of library extract region of interest

Identification Number	FullNameTH
NameEN	LastNameEN
BirthdayTH	BirthdayEN
Religion	Address
DateOfIssueTH	DateOfIssueEN
DateOfExpiryTH	DateOfExpiryEN

Recommend ⚠

• Image quality lowest should be 600x350
• Images with minimal reflections should be used. for good results
• Identity Card should be size in the image about 75%, if the image doesn't cropped that to be left only Identity Card area.
• For faster, please resize image and usage CUDA GPU.

Installation

Install using pip for stable release,

pip install thai-personal-card-extract

For latest development release,

pip install git+git://github.com/ggafiled/ThaiPersonalCardExtrac.git

Note 1: for Windows, please install tesseract first by following the official instruction here https://medium.com/@navapat.tpb/734dae2fb4d3 On medium website, be sure to setup already.

Note 2: for Linux os, please install tesseract by following the official instruction https://github.com/tesseract-ocr/tesseract

Usage

# With build-in Config Options.
import ThaiPersonalCardExtract as card
reader = card.PersonalCard(
    lang=card.THAI,
    provider=card.DEFAULT,
    tesseract_cmd="D:/Program Files/Tesseract-OCR/tesseract",
    save_extract_result=True,
    path_to_save="D:/dev/ThaiPersonalCardExtract/examples/extract")
result = reader.extractInfo('examples/card.jpg')
print(result)


# With free-style 
from ThaiPersonalCardExtract import PersonalCard
reader = PersonalCard(lang="mix", tesseract_cmd="D:/Program Files/Tesseract-OCR/tesseract") # for windows need to pass tesseract_cmd parameter to setup your tesseract command path.
result = reader.extractInfo('examples/card.jpg')
print(result)

# With free-style 
from ThaiPersonalCardExtract import DrivingLicense
reader = PersonalCard(lang="mix", tesseract_cmd="D:/Program Files/Tesseract-OCR/tesseract") # for windows need to pass tesseract_cmd parameter to setup your tesseract command path.
result = reader.extractInfo('examples/card.jpg')
print(result)

Output will be in list format, each item represents result of library can extract, respectively.

{
   "Identification_Number": "9999999999999",
   "FullNameTH": "นาย อายุมฺมุราเสะ",
   "PrefixTH": "นาย",
   "NameTH": "อายุมฺมุราเสะ",
   "LastNameTH": "อายุมฺมุราเสะ",
   "PrefixEN": "Me",
   "NameEN": "Shoys",
   "LastNameEN": "Hinata",
   "BirthdayTH": "21 มี.ย. 2539",
   "BirthdayEN": "21 Jun..1996",
   "Religion": "พุทธ",
   "Address": "ท๒ 99/1 มิชีโฮะ เขตฮานามิกาวา อำเภอชิบ;",
   "DateOfIssueTH": "11 ส.ค. 2554",
   "DateOfIssueEN": "~11 Ang. 2021",
   "DateOfExpiryTH": "11 ส.ค. 2574",
   "DateOfExpiryEN": "21 ug. 2092"
}

For set lang attribute to tha

from ThaiPersonalCardExtract import PersonalCard
reader = PersonalCard(lang="tha", tesseract_cmd="D:/Program Files/Tesseract-OCR/tesseract") # for windows need to pass tesseract_cmd parameter to setup your tesseract command path.
result = reader.extractInfo('examples/card.jpg')
print(result)

Output will be in list format, each item represents result of library can extract, respectively.

{
   "Identification_Number": "9999999999999",
   "FullNameTH": "นาย อายุมฺมุราเสะ",
   "PrefixTH": "นาย",
   "NameTH": "อายุมฺมุราเสะ",
   "LastNameTH": "อายุมฺมุราเสะ",
   "BirthdayTH": "21 มี.ย. 2539",
   "Religion": "พุทธ",
   "Address": "ท๒ 99/1 มิชีโฮะ เขตฮานามิกาวา อำเภอชิบ;",
   "DateOfIssueTH": "11 ส.ค. 2554",
   "DateOfExpiryTH": "11 ส.ค. 2574"
}

And you can set ocr provider following below default #used both easyocr and tesseract **Recommend Or easyocr Or tesseract

from ThaiPersonalCardExtract import PersonalCard
reader = PersonalCard(lang="tha", provider="default", tesseract_cmd="D:/Program Files/Tesseract-OCR/tesseract") # for windows need to pass tesseract_cmd parameter to setup your tesseract command path.
result = reader.extractInfo('examples/card.jpg')
print(result)

Config Options

you can set options to Instance by below keyword

Parameter name	Value Type	Example
lang	String	Expected Results Language bash mix #get all area both tha and eng Or bash tha Or bash eng *Default is 'mix'
provider	String	OCR Provider have bash default #used both easyocr and tesseract **Recommend Or bash easyocr Or bash tesseract *Default is 'default'
template_threshold	Double	Rate to cals similarity of template *Default is 0.7
sift_rate	Int	Feature Keypoint rate *Default is 25,000
tesseract_cmd	String	Path of your tesseract command **For windows only.
save_extract_result	Boolean	Set True if you want to save extracted image *Default is False
path_to_save	String	Path that you given it save extracted image, relative with save_extract_result=True

Donate Me ☕

Mr.Nattapol Krobklang