/CNIC-Forms-OCR-detection-Text-Detection-Extraction-using-Deep-Learning

Primary LanguagePython

CNIC Optical Character Recognition

Problem Statement

Detect Name, Father Name, Gender, Date of birth, Date of Issue and Date of expiry from CNIC.

Solution

We discover the solution of this problem using easyocr python library. easyocr takes an image as input and calculate the bounding boxes of text.

Installation

pip install -r requirements.txt

This requirements.txt file includes all necessary libraries to run this project.

Code Exmaple

import easyocr
reader = easyocr.Reader(['en'])
image_path = "data/cnic_5.jpg"
result = reader.readtext(image_path)

Sample_CNIC

Results

image

you can get insights from graphs that your image size and image DPI(Dots Per Inch) is good then accuracy will be 100%.

Deployment

We are using latest technology called Docker to deploy this app. Following are the steps and instructions before deployment:

(1) Download and Install Docker for windows from Docker
(2) Create a Dockerfile in Visual Studio Code using following code

# use python as base image
FROM python:3.7

RUN pip install virtualenv 
ENV VIRTUALENV=/env
RUN virtualenv venv -p python3
ENV PATH="VIRTUAL_ENV/bin:$PATH"

# RUN apt-get update ##[edited]
# RUN apt-get update
WORKDIR /app
COPY . /app
# RUN pip install skia-python==86.0
# RUN apt-get install libgl1-mesa-glx-lts-utopic libgl1-mesa-dri-lts-utopic
RUN apt-get update -y 
RUN apt install -y libsm6 libxext6
RUN apt update
RUN pip install pyglview
RUN apt install -y libgl1-mesa-glx
RUN pip install -r requirement.txt

# Expose Port
# EXPOSE 5000

#Run the application
# ENTRYPOINT python app.py
CMD ["python","app.py"]
# CMD ["flask", "run", "-h", "0.0.0.0", "-p", "5000"]

(3) Create the Docker Image from Dockerfile using following commands

## docker build image syntax
docker build -t <image-name>:<version> .

## for example
docker build -t cnic_extraction:latest .

(4) Save the docker image for the future use using following command

## docker save image syntax

docker save -o <path for generated tar file> <image name>

## for example
docker save -o \docker_images\cnic_extraction.tar cnic_extraction:latest

(5) Load the docker image using following command

## docker load image syntax

docker load -i <path for generated tar file>

## for example
docker load -i \docker_images\cnic_extraction.tar

(6) Run Container from docker image using following command

## docker container run syntax
docker run -p <app_port>:<docker_map_port> <image_name>

## for example
docker run -p 8080:8080 cnic_extraction

(7) it will run flask app at localhost:8080 in your default browser like this:

Input

image

Output

image