How to extract text from images using EasyOCR Python Library (Deep Learning)

[@rakpongkittinaradorn7202]

Hi, thanks for making this educational content. EasyOCR is now supporting Devanagari (Hindi, Marathi, Nepali). Other languages are also coming.

[@bhattbhavesh91]

Thanks for sharing

[@KenJee_ds]

This is so cool! Great video Bhavesh!

[@bhattbhavesh91]

Thank you so much Ken :)

[@maverick6912]

This is exactly something I needed for a hackathon. Thanks a lot.

[@bhattbhavesh91]

Welcome!

[@rahulchavanrc8418]

​@@bhattbhavesh91​ hello sir plz upload Inovice in Ocr vidio

[@girishpradnya]

Thanks Bhavesh for amazing video. Please suggest any pre processing steps required for natural images. Because most of the ocr have not given me accurate results if images are directly passed without any pre processing.

[@tinky5418]

Thanks for sharing. I found that the effect of Arabic is not good enough in EasyOCR. Have you used PaddleOCR? Can you share it?

[@onlineaccadmy9659]

i enter the command ...... reader=easyocr.Reader(['en'])....and then i get the messege CUDA not available - defaulting to CPU. Note: This module is much faster with a GPU. plz solve this

[@rashasalim1259]

Thanks a lot for this great video! :) You mentioned that it is possible to use it on mobile devices also. I'm trying to do just that, but couldn't find any resources should I download the model from pypi and use that? Any help or advice is highly appreciated

[@kamalakannandurai9153]

Hi, Thanks for sharing. Can we use the text extracted from the image as input for other purpose in the same program.

[@Joel95ify]

This is better than tessersact! Thank you for the video!

[@bhattbhavesh91]

Yes! its amazing! I'm glad you liked the video!

[@ayanmalik926]

Hyy bro can you say how to find the coordinates of bounding boxes

[@unknowndeveloper4495]

what was the last part of cv2.rectangle(image,(x_min, y_min), (x_max, y_max)(0,0,255),2) the (0,0,255),2) ?

[@samzion4974]

what is the platform you are using for coding and testing ? kindly tell me about it . 😀

[@ChrisCastaneda-k7e]

I'm new to this concept. How can I run this in my command terminal? Do I need to save this as a .py file? What are the steps to be able to do this. Thanks

[@tchagonda]

is it possible to supply coordinates and extract text only from that bounding box

[@jyotiahire1240]

Thanks for sharing, but i have question can we get text ast it is after extraction? for example i need movie title as it is after extarction.

[@Gabriel-iy1zq]

Great Video, something new for once! I had trouble although regarding the PIL module. I do not know why but when I run the code it outputs an error saying: File "C:\Users\angel\PycharmProjects\ImageProc\venv\lib\site-packages\easyocr\detection.py", line 4, in from PIL import Image ModuleNotFoundError: No module named 'PIL' I do not understand this since I already have Pillow installed. I do not know Why this is happening. Can someone help?

[@chinmayar4202]

when i run reader=easyocr.Reader(['en']) command , i am getting error as CUDA not available in google collab. Can you please tell how to solve this error?

[@Bizet1985]

Go to Runtime > Change Runtime Type > choose GPU

[@williamflores7323]

Is ir better than tesseract?

[@shyamalaspure3316]

How can we load again model once it is saved in memory. I am trying to use easyocr.Reader(['en'], gpu = False) It is again downloading it. I have seen my local it is saved as .pth file.

[@lucasaraujo6.]

Hi, great video!!! I'm using images with vertical bars ( | ) separating some numbers and the program is recognizing the simbols as dashs ( / ). I would like to get only the numbers. Does there have any filter to take the bars out of my treatment? Is it possible to get only the numbers? Thank you a lot fot the content!

[@dungama]

Hi Bhavesh.Nice video.Why don't you start affordable DS or machine learning video in a sequential manner.Even after 2 or 3 days,you can upload the same videos to youtube also.

[@bhattbhavesh91]

I do this as a hobby! I haven't thought of creating an end to end course!

[@PrakashKing8342]

Do you have solutions for alpha numeric recognition I have tried all the method finally I’m getting fail tesseract easy ocr etc for example ABC00OO1 and SI1234H

[@gauravranchi]

I was trying to find a software like this which can translate doctors prescription...do you know where to find it or can you help me develop one .

[@dhananjaykansal8097]

Your videos are so interesting.....

[@bhattbhavesh91]

Glad you think so!

[@ferozmohammad5841]

Can I use easyocr to extract text from pdf?

[@bhattbhavesh91]

Sure! you can extract the image from the pdf and then run easyocr on it!

[@bhupeshpattanaik7150]

will it work if image has hand written text ?? please answer

[@deepthik6828]

Good.It working nicely

[@bhattbhavesh91]

Great

[@puneetbhushan5536]

Thanks for sharing, does the library also recognizes numeric character in an image?

[@bhattbhavesh91]

Yes it does!

[@ayanmalik926]

Hyy bro how did your write the coordinates of bounding boxes

[@shivampm]

How I can loop through all detected texts and plot mark on single file?

[@elifozcan7571]

Please help.. I did not understand the idea in this code: cord = output[-1][0] --> It allows me to choose machine text but I did not understand this point. In python normally index numbers start from 0 .. what does it mean -1 ? Please guys help me. Thx a lot

[@TrueBlueProd]

Is there a way to execute this in Visual Studio Code?

[@fabrizio-6172]

Hello there is any way to manage it stating from a PDF?

[@nikitha_krishnan]

Getting "Unknown C++ exception from OpenCV code" error. PLEASE REply

[@deepak464]

Hi Bhavesh, Is there any package that you know for language translation from Swedish to English etc?

[@koppuladattatreya1623]

Can we print the text using bounding box, and can we compare of two texts from two different images using bounding boxes?

[@vishnuvasansrinivasan7797]

i need the model which is imported in 'reader' variable

[@GuruTechHub]

hi. please make video on extract hindi table contains text in devnagri or utf-8 to csv from images. i try lot on inter but not found any video or method.. please make video on this it will help lot

[@arunadang7872]

Is it jupyter notebook? If not which CLI is this?

[@bhattbhavesh91]

Yes, its jupyter on Google Colab!

[@abieltesfamichael5210]

Hi sir i am working in developing this project on our local language Tigrigna which is Eritrean language..can you help me how to prepare the language data please?

[@sagihaviv5675]

sir is there a way to make a recognition of a word/line of text, if it is a code or not.

[@shishirtimilsena654]

Can we deploy this ??? if so, can you please give an idea ??

[@bhattbhavesh91]

Yes! Sure, I'll create a video on this soon!

[@shishirtimilsena654]

@@bhattbhavesh91 great...then waiting for it

[@rajuvarma-zk2zw]

Please do on Hand written text extraction

[@mahanandadimulam5829]

Sir can we scan the image from the camera and then convert it to text if we can how?

[@adityaghosh8601]

Is this better than tesseract by Google.

[@harsheenrajpal659]

Yes, I tried tesseract first and then cam across this later, works MUCH better

[@misbakhulislam8773]

can i use this on my website?

[@raghubayya3568]

thank you for amazing video, i have quick question after extracting text from image is it possible to store output in excel or word doc. i will be glad

[@bhattbhavesh91]

Yes you can!

[@raghubayya3568]

@@bhattbhavesh91 How about reading from pdf file ?

[@MrHamza0001]

@@raghubayya3568 you need to save the pages of pdf file/s as 'jpg' and then run OCR over them. You can do it by using 'wand' library

[@shashwatvishnu3109]

how can we remove that float value in output

[@ismailidowu3167]

Can this work on other IDE like pycharm and jupyter notebook?

[@bhattbhavesh91]

This is Jupyter Notebook on Google Colab :)

[@carlcontreras8963]

AttributeError: module 'torch.backends' has no attribute 'mps'

[@sankarachari]

I have tried in CMD getting error 'ModuleNotFoundError: No module named 'PIL' '

[@BorNation]

pip install pillow

[@RajaSekharGowda]

nice, but it gives low accuracy for old images...

[@parthchokhra7298]

Great Library !!

[@bhattbhavesh91]

Yes! Really great work done by the creators!

[@qamarzaman4390]

can you share this source code and which algorithm is used for recognition please

[@santhoshrangasamy528]

Useful.

[@bhattbhavesh91]

Glad it was helpful!

[@attamuhammad7240]

I am looking for OCR which could read blank spaces as well, can google vision api do this ?

[@ahmedhazem3986]

hey bro do u found something helped u i need help too

[@attamuhammad7240]

@@ahmedhazem3986 I have done this project. You could contact me on my what's app number : 0336-6463338

[@ShahidulIslam_personal]

i want to print only output of text.

[@aravindeswarkotni]

Hello sir I have a doubt I am working on a deep learning project In that project i have to extract Brand name of the products based on image How can I extract it Can you please help me sir

[@codingandgamesforfun2234]

Hey aravindeshwar. Are you still making projects?

[@elbi137]

Does this library works for 'Devnagari' ?

[@bhattbhavesh91]

Not at this stage!

[@narjesatia]

Thanks for sharing, is there Arabic language please ?

[@bhattbhavesh91]

Welcome! I think Easyocr has support for Arabic language! Do check the official repository.

[@narjesatia]

@@bhattbhavesh91 i will check..thank u so much

[@santhoshguggilla4470]

awsome

[@bhattbhavesh91]

Thank you!

[@soumyaranjansethi1790]

Nice👌👌

[@bhattbhavesh91]

Thanks 😊

[@RAZZKIRAN]

will it work for hand written text ?

[@bhattbhavesh91]

No!

[@RAZZKIRAN]

@@bhattbhavesh91 how to recognize from hand written text, i sent mail to u

[@bhattbhavesh91]

You will have to train a custom model for it

[@RAZZKIRAN]

@@bhattbhavesh91 help me in coding

[@wariszaya8089]

Does it extract handwritten text?

[@bhattbhavesh91]

I don't think so!

[@hemantsah8567]

@@bhattbhavesh91 Do you know any library which works for handwritten notes, if not what are other options?

[@mohammadriduan3676]

Is it support bangla language??????

[@bhattbhavesh91]

Yes, read more here - www.jaided.ai/easyocr/

[@himanshuparashar5883]

How to use hindi word please reply ☹️🥺🥺

[@bhattbhavesh91]

I have 1 more video to detect hindi characters. Do search the video section of my channel.

[@huakai_1497]

kernel keep dying when import easyocr

[@bhattbhavesh91]

try running it on Google Colab!

[@123-l3k]

extract text from text?

[@sankarachari]

Hi This is good. Can we run this without GPU?

[@bhattbhavesh91]

Yes, you can but the prediction will be slow!

[@sankarachari]

@@bhattbhavesh91 I wanted to do some experiment like Google lense

[@sankarachari]

Can you build a model with deep learning which Will extract text from image?

[@sankarachari]

I need it in Python

[@elifozcan7571]

Please help.. I did not understand the idea in this code: cord = output[-1][0] --> It allows me to choose machine text but I did not understand this point. In python normally index numbers start from 0 .. what does it mean -1 ? Please guys help me. Thx a lot

[@codingandgamesforfun2234]

Python supports negative indexes also. So -1 means that you are approaching the last element, similarly -2 means indicating second last element and so on.