letter detection in opencv python

Question

I want to detect the letters in the image of the following detective's answer.source here.

I answered it with kanren logic code:

from kanren import run, eq, conde, var, Relation, facts
from kanren import run, var, fact, Relation

x = var()

murderer = Relation()
# Murder suspects
facts(murderer ,("Mr Ronald", "Jack Green"),
           ("Mr Ronald", "John jacobson"),
           ("Mr Ronald", "June Green"))

sol = run(3,x,murderer("Mr Ronald",x))


findings = Relation()

facts(findings,("First of","January"),
        ("Fourth of","October"),
        ("Fifth of","March"),
        ("Third of","June"))

January = run(1,x,findings ("First of",x))
October = run(1,x,findings("Fourth of",x))
March = run(1,x,findings ("Fifth of",x))
June = run(1,x, findings("Third of",x))

print(January[0][0])
print(October[0][3])
print(March[0][4])
print(June[0][2])

So, Kanren's logic answered that the culprit was: John

So when I want to detect letters that match opencv it fails and keeps on failing. I study here and here.

I want the output to be like , and it matches the answer from kanren's logic which is: John

Thank you, any help will be highly appreciated.

original

all the "kanren" stuff is irrelevant to "detecting letters", which is called Optical Character Recognition (OCR). are you asking to be recommended an OCR library? please take the tour and review How to Ask and minimal reproducible example. — Christoph Rackwitz
– Christoph Rackwitz, Commented Oct 25, 2022 at 7:59

Yunus Temurlenk · Accepted Answer · 2022-10-25 06:19:47Z

1

EasyOCR can be a good choice of reading this kind of texts. Its accurate and give you the position of each letter in the text image.

When I tried the code below with your image:

import easyocr
reader = easyocr.Reader(['ch_sim','en'])
result = reader.readtext('/ur/img/dir/img.jpg')
print(result)

I get the output:

[([[190, 99], [525, 99], [525, 156], [190, 156]], 'Firstof Jahuary', 0.8845715173000117),
([[173, 152], [544, 152], [544, 204], [173, 204]], 'Fburthof October', 0.8176276380919366),
([[212, 208], [504, 208], [504, 258], [212, 258]], 'Fifthof March', 0.7889743493622021),
([[224, 264], [494, 264], [494, 314], [224, 314]], 'Thirdof Juhe', 0.8849983221023138)]

I can say it is a %95 accurate result.

answered Oct 25, 2022 at 6:19

Yunus Temurlenk

4,4154 gold badges21 silver badges45 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

letter detection in opencv python

1 Answer 1

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related