Python code for marking regions of interest in an image for OCR

Marking_ROI.py

# use this command to install open cv2
# pip install opencv-python

# use this command to install PIL
# pip install Pillow

import cv2
from PIL import Image

def mark_region(imagE_path):
    
    im = cv2.imread(image_path)

    gray = cv2.cvtColor(im, cv2.COLOR_BGR2GRAY)
    blur = cv2.GaussianBlur(gray, (9,9), 0)
    thresh = cv2.adaptiveThreshold(blur,255,cv2.ADAPTIVE_THRESH_GAUSSIAN_C, cv2.THRESH_BINARY_INV,11,30)

    # Dilate to combine adjacent text contours
    kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (9,9))
    dilate = cv2.dilate(thresh, kernel, iterations=4)

    # Find contours, highlight text areas, and extract ROIs
    cnts = cv2.findContours(dilate, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
    cnts = cnts[0] if len(cnts) == 2 else cnts[1]

    line_items_coordinates = []
    for c in cnts:
        area = cv2.contourArea(c)
        x,y,w,h = cv2.boundingRect(c)

        if y >= 600 and x <= 1000:
            if area > 10000:
                image = cv2.rectangle(im, (x,y), (2200, y+h), color=(255,0,255), thickness=3)
                line_items_coordinates.append([(x,y), (2200, y+h)])

        if y >= 2400 and x<= 2000:
            image = cv2.rectangle(im, (x,y), (2200, y+h), color=(255,0,255), thickness=3)
            line_items_coordinates.append([(x,y), (2200, y+h)])


    return image, line_items_coordinates

Dear @akash-ch2812 or @jlumbroso,

I'm having some trouble using this code on some PDFs I'm working on, I thought either you might have some ideas on what's going wrong.

I have a single PDF called, "DNR_WFH.pdf". I was able to use the first section of code below to split the PDF into separate pages and name them accordingly.

from pdf2image import convert_from_path

pdfs = r"C:\Users\mhiebing\Desktop\DNR_WFH.pdf"
pages = convert_from_path(pdfs, 350)

i = 1
for page in pages:
    image_name = "Page_" + str(i) + ".jpg"
    page.save(image_name, "JPEG")
    i = i+1

For the second section, I'm only looking at the first JPEG to keep things simple. When I call the mark_region(image_path) function nothing happens. Is there supposed to be a window where I outline the boxes we want to extract text from? Here's what I have for the second section:

import cv2
import matplotlib.pyplot as plt


def mark_region(image_path):

    #define the mark_region method
    FILENAME = r"C:\Users\mhiebing\Documents\GitHub_Repos\MonthlyStatsExtract\Page_1.jpg" # <--- change this to be the file you want
    image, line_items_coordinates = mark_region(FILENAME)
    plt.figure(figsize=(20,20))
    plt.imshow(image)
    plt.savefig("image-with-regions.png") # <--- added this to output an image

    image = cv2.imread(image_path)

    # define threshold of regions to ignore
    THRESHOLD_REGION_IGNORE = 40

    gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
    blur = cv2.GaussianBlur(gray, (9,9), 0)
    thresh = cv2.adaptiveThreshold(blur,255,cv2.ADAPTIVE_THRESH_GAUSSIAN_C, cv2.THRESH_BINARY_INV,11,30)

    # Dilate to combine adjacent text contours
    kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (9,9))
    dilate = cv2.dilate(thresh, kernel, iterations=4)

    # Find contours, highlight text areas, and extract ROIs
    cnts = cv2.findContours(dilate, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
    cnts = cnts[0] if len(cnts) == 2 else cnts[1]

    line_items_coordinates = []
    for c in cnts:
        area = cv2.contourArea(c)
        x, y, w, h = cv2.boundingRect(c)

        if w < THRESHOLD_REGION_IGNORE or h < THRESHOLD_REGION_IGNORE:
            continue

        image = cv2.rectangle(image, (x,y), (x+w, y+h), color=(255,0,255), thickness=3)
        line_items_coordinates.append([(x,y), (x+w, y+h)])

    return image, line_items_coordinates

mark_region

Thank you for putting up the article and supplying the code!

@Matthew-Hiebing
I'm just trying to adapt the existing fantastic code and I found your question.

This code snippet of your code is inside the function mark_region but should be outside. This because you are calling the function mark_region(FILENAME) and pass a FILENAME:

    #define the mark_region method
    FILENAME = r"C:\Users\mhiebing\Documents\GitHub_Repos\MonthlyStatsExtract\Page_1.jpg" # <--- change this to be the file you want
    image, line_items_coordinates = mark_region(FILENAME)
    plt.figure(figsize=(20,20))
    plt.imshow(image)
    plt.savefig("image-with-regions.png") # <--- added this to output an image

Thank you so much for your work.
There were problems with opencv-python-4.5.5.62.
Use opencv-python==4.1.2.30.

if w < THRESHOLD_REGION_IGNORE or h < THRESHOLD_REGION_IGNORE:
            continue
        
        image = cv2.rectangle(image, (x,y), (x+w, y+h), color=(255,0,255), thickness=3)
        line_items_coordinates.append([(x,y), (x+w, y+h)])

Thanks a lot man, Before I was only getting few area marked but after revamp and using your code it's marking every single area
Thank you once again. Keep up the good work

if we want to optimize this code:

import cv2
from PIL import Image
from pdf2image import convert_from_path

def mark_region(im):
gray = cv2.cvtColor(im, cv2.COLOR_BGR2GRAY)
blur = cv2.GaussianBlur(gray, (9,9), 0)
thresh = cv2.adaptiveThreshold(blur, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, cv2.THRESH_BINARY_INV, 11, 30)
kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (9,9))
dilate = cv2.dilate(thresh, kernel, iterations=4)
cnts = cv2.findContours(dilate, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)[0]
line_items_coordinates = []
for c in cnts:
area = cv2.contourArea(c)
x, y, w, h = cv2.boundingRect(c)
if y >= 600 and x <= 1000 and area > 10000:
image = cv2.rectangle(im, (x,y), (2200, y+h), color=(255,0,255), thickness=3)
line_items_coordinates.append([(x,y), (2200, y+h)])
elif y >= 2400 and x<= 2000:
image = cv2.rectangle(im, (x,y), (2200, y+h), color=(255,0,255), thickness=3)
line_items_coordinates.append([(x,y), (2200, y+h)])
return image, line_items_coordinates

poppler_path = r 'you should write poppler bin folder path '
pages = convert_from_path("you should write here pdf path", 480, poppler_path=poppler_path)

for i, page in enumerate(pages):
image_name = f"Deneme_{i}.JPEG"
page.save(image_name, "JPEG")
im = cv2.imread(image_name)
marked_image, coordinates = mark_region(im)
cv2.imwrite(f"Marked_{image_name}", marked_image)