Text recognition (OCR) with Tesseract and Python

[@ryanmoye9189]

One issue with the Chinese text is that the first two characters are in traditional Chinese and the second two characters are the simplified. So the first time you ran it and it gave you 汉语，汉语 was correct as it converted the traditional into simplified because you used chi_sim.

[@elieferland8561]

Thank you for all your good tutorials! Could you make a video on natural scene text detection using opencv and EAST one day??

[@michaelchung9119]

This is a really good tutorial. Thank you

[@chakrabmonoj]

Hi - thanks for the excellent tutorial. This really helps a lot. While running your code on the image that I have extracted from my PDF document, instead of printing the text in the image file, it is spitting out some texts as below : code : image =cv2.imread("page-12.png") cv2.waitKey(0) Output : -1 code : text = pytesseract.image_to_string(image) Print(text) output : DF Metadata Extraction with Python file = PyPDr2 xnp = pdf_file.getimphetadatal dict = 0 4 in mp methods xnp_dict (i) xnp_dict (i) pp.pprint xmp dict) anytestringobject if pyPDr2 is unable to decode the string (Fenniak, 2016b). Putting these two methods together yields a custom function that can be used to extract document information dictionary metadata from PDF files (sce igure 9). The resulting DocumentInformation object which generated by the custom get_doc_info() function contains a dictionary with five key:value pairs (see Figure 10). This extracted data ‘matches the raw metadata located in the Document Information Dictionary abject located at 208 0 obj inthe file (see Figure 3). (© 2019 The SANS institute Author retains ful ight Am I getting something wrong? Unfortunately I am unable to show you how the input looks. Any help much appreciated.

[@chakrabmonoj]

Hi...I ran your code again and it is working. So you may ignore the above comment. The problem is my file has a dark background and so with the values that you have shown for adjusted threshold do not seem to be converting the background to white and the print(text) is still not showing the text embedded within the picture. My picture file is PNG - will that require a different adjustment to your code and the threshold values? Thanks - really appreciate

[@ForMyOwn_1]

It was very informative and helpful lesson. Thanks

[@marwen2594]

thanks for the tutorial i have question please , can tesseract ocr detect handwriting if so can you make another tutorial about that

[@Benm2793]

Adaptive threshold is a great tip. Thanks!

[@wimr.9672]

thanks for the tut! Helped me a lot

[@ahmedhelal920]

Very good introduction to ocr . thanks 😊

[@RapidView]

Tnx a lot. What would be the processing when u get dynamic images.

[@hemantchauhan6437]

NEED HELP! I am making a website where user can upload a pdf but I want that pdf to upload only if that pdf has images of only HANDWRITTEN text. Thank you for reading.

[@davidralte4572]

Thank you for Your help, May God Bles You.

[@sauravsinha8746]

Will you tell how we can get this data in csv format

[@heinhane]

This is very helpful . Tanks

[@bibhutirajansingh]

Can a multi-page PDF be OCRed this way?

[@ok-kp5jn]

Great content!!

[@jay4866]

Hi can you do the same thing using reberry pi or Arduino.

[@len5204]

Hi, thanks for the tutorial. But i was wondering how it will be if there's gonna be an upload photo feature. So we dont have to change the image filename to be used everytime. Is it possible?

[@maloukemallouke9735]

Thank you for sharing, I am wondering if OCR it's a heavy process to find digits on large images?

[@AI_CANISTER]

its simple, to get only digits in a large image,, config = r'--oem 3 --psm 6 outputbase digits' digitd = pytesseract.image_to_data(img, config) when you print digits you get only digits

[@maloukemallouke9735]

@@AI_CANISTER thanks but i already tested not work

[@tushargawade2045]

really Great !!! and helpful. if we train our model using cnn , will it increase accuracy?

[@sauravsinha8746]

How can we save the output in csv format please

[@dimitheodoro]

Thanx alot!!!

[@firegames2741]

Thank for such a useful video. I need help from you, can you convert captcha file to text. I'm trying, but not converting properly.

[@andreadotta6653]

Hi Sergio, Thanks for the video! I've a question. Can you have something ideas to convert raster images in vector images (es. jpg to svg)?

[@eksimedya4659]

Danke schön mein bruder

[@marienoellevandervlugt9183]

I love tour tuto, i'am trench, englich it's difficile. But i want learn python

[@r-beanmondy6203]

if I wanna use it on live video which part that I should change for my code?

[@gawaderajesh]

I am gettng below error... Please help raise TesseractError(proc.returncode, get_errors(error_string)) pytesseract.pytesseract.TesseractError: (1, 'Error opening data file C:\\Program Files\\Tesseract-OCR/tessdata/chi_sim.traineddata Please make sure the TESSDATA_PREFIX environment variable is set to your "tessdata" directory. Failed loading language \'chi_sim\' Tesseract couldn\'t load any languages! Could not initialize tesseract.')

[@gowthamns8228]

Wow very good, but the problem is "If the text is very clear and crisp its is showing output correctly", But I want to know if the image has multiple data not only text for example "bills, taken calendar photo or any kind of images", How to print the string from that, I tried my self it's not printing anything, Any idea for this?

[@AI_CANISTER]

it can detect digits since date is mostly in digits i think it will work well. config = r'--oem 3 --psm 6 outputbase digits' digitd = pytesseract.image_to_data(img, config) when you print digits you get only digits

[@vinsmokearifka]

thank you. how to get only title?

[@sauravsinha8746]

How can we get this data into csv format

[@riyajagtap5006]

CAN IT WORK FOR 7 SEGMENT LED

[@opendllmaster5125]

Isn't it possible to train Tesseract to improve the reading?

[@abdelrhmanshokr7546]

this helped a lot but still there is a date value that it doesn't seem to get it I don't know why to be honest

[@user-zo2gm1bh1f]

Please sir lang arbic text = pytesseract.image_to_string(adaptive_threshold, config=config, lang "arbic" ) ARB OR AR OR What ???

[@sauravsinha8746]

Can we save to Csv format

[@dukeofminecraft]

can you train pytesseract on handwriting data and return the string data ?

[@ashishzarekar9599]

could you please help on how to implement for scanned and digital pdf?

[@lakshmitejaswi7832]

Make a video on how to build custom ocr

[@nikolaydd6219]

Thanks

[@LuisGarcia-tb9po]

Tesseract has been adding arrows to each each cell in my excel spreadsheets, anyone know why that might be? It recognizes every word and number correctly but adds some kind of ‘illegal” character code that is excel displays as arrows then boxes with a question mark inside

[@thesupremeeagle2146]

Do you have another way to use tesseract. I want to turn my programme into executable for people who want to download it, but to make it work they need this tesseract-ocr file. So i put it into the download file but the tesseract file is too heavy ! I dont want to make my programme at 1GO because of tesseract ! Please help :c

[@AKKJ420]

Do an ANPR mate

[@juanpajaro4084]

Gracias amigo, me resolviste muchas dudas.

[@senpaikun5947]

hey... im not able to print that chinese letters in my output... can asome one help me oue plz

[@luismata4086]

How can I make than the algoritm recognize Latex language? ---> pytesseract.image_to_string(img, lang = '?') . What have to use for the parameter "lang"?

[@NicolaMastrandrea]

Grazie 😊

[@revudevendraswamy6632]

Is it work for handwritten data ??

[@kascesar]

hello, im looking for a rcnn for this task, do u know a nice one for this task ?

[@lukmanchaiyarab1451]

can you please show digit recognition ;thank in advance

[@AI_CANISTER]

config = r'--oem 3 --psm 6 outputbase digits' digitd = pytesseract.image_to_data(img, config) when you print digits you get only digits

[@Terminator-lx5jx]

You didnt solve the text though after pre processing

[@rajeevkalaskar6373]

I want to make a desktop application for this. how can I do it. Need help 🆘

[@pysource-com]

For commercial projects/consulting services you can contact me here: pysource.com/services

[@towhidurrahman8202]

is this possible for number plate recognition form this code ? and the language in bengal

[@AI_CANISTER]

It can't recognize number plate but it can extract the digits and alphabet after detecting the number plate with a different method

[@ankursoni8060]

How to detect a text from a particular co-ordinate of an image?

[@pysource-com]

First you need to cut that region. Check my youtube videos "Crop images" and you'll know hot to d that. once you did cut the portion, you can parse that one to the OCR

[@kisamesafe]

it could be shorter

[@sarthakgarg6531]

how we can read different font like it the image has italic font so how can we do that ?

[@AI_CANISTER]

I've tried the same method and it worked, I'm sure it will work for you too

[@jeffu73]

For me, the image is not getting recognized.

[@raquelcosta2730]

same for me :(, i have tried different clear images and it's not working. Any tips?

[@AneleMbabela]

Instructions and source code link is broken.

[@pysource-com]

I've just fixed it, thanks for pointing that out

[@AneleMbabela]

@@pysource-com Thanks for the work you've been putting out. Its really making a difference. God bless you, brother..

[@NoamHarel-Google-Is-The-Best]

need some help.. it's wirte this line... You need configured Python 2 SDK to render Epydoc docstrings thanks a lot

[@nuwanthajayasinghe115]

I installed tesseract and try to work in vscode.But when programming python from vscode, tesseract could not be imported into vscode.Can you tell me how to import tesseract to the vscode??

[@jeffu73]

what is he using vscode or?

[@rohitnara6738]

cv2.imread("img.gif") is not working how can we read text from .gif file type please tell

[@jasonlo3429]

The Chinese words are traditional and not simplified. Change the language to chi_tra and it should work better

[@aichamahfoudh2451]

How can we use this on google colab?