Efficient Text Detection & Recognition Tutorial | EAST & Tesseract | Python OpenCV-with code

This is a quick tutorial of how we can use the EAST Text Detection and Text Recognition tools. Project is developed on Python and OpenCV
Support me on Patreon

💻 General Tech

📸 Camera Gear

🎥Video Accessories

👨‍💻Books I like on Personal Growth

👨‍💻Books I like on Investing

DISCLOSURE: Some of the links on this page are affiliate links, meaning, at no additional cost to you, I may earn a commission if you click through and make a purchase. Affiliate commissions help fund videos like this one.

image to text conversion

[hemantchauhan]

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.

[Gabriel-iyzq]

I have a problem that it identifies some text my images but it prints them out on the image with additional question marks at the end. For example, if the text on the image is "Hello World", it would print "Hello??" "World??". I tried replacing the new line elements the string may have but that only took care of one of the question marks. Any idea ho to take care of the other ones (sometimes is more than one, depending on the image).

[nimeshkumar]

Hi, I am trying your code for text detection. but I want to prune this frozen model, how can I prune this model. OR is this model is already pruned.

or any other way I can shrink the size of this east model, I have tried quantization and that brings the model size from 96MB to 24MB, I want to shrink further say 10MB.

Can you please help me

[nimeshkumar]

Hi, i just need to see is the text is present or not, so can you please help me which lines i have to remove

[mihirkarnik]

how tp use this code in google colab?
its showing ModuleNotFoundError on import pytesseract, can you help with that

[jaychavda]

is there any difference from the code explained in the video and the code uploaded on github?
because I am running the same exact code and the text is detected in all the images but while showing the text it shows only "?" question marks for certain words like "self" and "isolate" as well as "subscribe like and comment " . Is there any solution to this?

[kirankumarchincholi]

Thank you so much you saved my lot of time, I run you code in my system it working fine but there are some extra garbage characters added in the text like ??, How can I remove that?

[ariouathanane]

Hello, thank you very much for all. I executed the code but i have only the boxes but not the recognition of the textes.. i check the code in your video and the code in github, it exactly the same but i don't know why i don't have the recognition of the textes

[yvijayashri]

can you please share the code for text detection (east) and recognition(tesseract) for video as input through webcam. can you please say how to do that

[harshanb]

Can I apply for video instead of images

[SAINIVEDH]

How to do custom training for EAST ?!!

[anburenganathan]

Thanks for the video misbab
Am trying to convert gif image into text but am getting only 40% accuracy becoz the font style on the image is recursive.. pls help me to resolve this.. if you want a sample image i will share with you

[xilluminati]

Thank you for clearly explaining each line!

[alokprasad]

Hey I am working on developing a text detection model for registration certificate which carry alphanumeric terms. Can this be beneficial for the same?

[manjubadiger]

Hi wonderful video thanks so much, can I apply your tutorial and code for invoice image

[ahmedumermughalradiantscho]

Bro can you make a tutorial on how to upload an image an google or facebook or insta ai doesn't know whats in the picture..

[Gabriel-iyzq]

Great video. This is something you don't see every day on youtube.
I have a question although about the cv2.cvtColor function used. It gives me an error:
cv2.error: OpenCV(4.4.0) error: (-215:Assertion failed) !_src.empty() in function 'cv::cvtColor'

Just so you can know what I have done, I have check the coordinates of startX, startY, endX, and endY and they at least have values. After that, I verified the roi by displaying it with cv2.imshow() and got an image very small that appears to have no text at all in it. I have check my code various times and don't find any differences.
Any help?

[mirrorZ]

It's amazing sir
I expecting emojis detection and removal plz make a vedio on emojis.

[antonioranielcayaco]

Nice video bro. Im having a problem on improving the image so that pytesseract will be able to read it more accurately. Cause the letters are too close to each other.

[naveedakhtarbutt]

VERY HELPING TUTORIAL . Can we detect urdu language text from same methodoly?