OpenCV 3 KNN Character Recognition Python

GitHub site:

Prerequisite:

If you found this video helpful please consider supporting me on Patreon:

[ВладимирАлександров-ог]

if shift recognises as key, add or replace values from intValidChars variable to lowercase keys values

[DinukDeshan]

For the people with Shift key issue, keep shift pressed when you run GenData.py. Then release at numbers. I tried keeping Capslock on but it did not give correct capital letter ASCII values. Great video

[TungNguyenThanh-pbvy]

Excuse me!Do you know what the files flattened_images.txt and classifications.txt mean? What is their meaning?

[indiansoftwareengineer]

thanks for helpful video tutorial. but I am getting output only as
what I am missing here?
please explain steps @5:15??

[lchunleo]

Thanks for sharing. It is informative. Does the generation of the dataset works for SVM classifier?

[missionx]

+Chris Dahms Hello, while i was running the GenData file i got the classification.txt and flattened_images.txt files with "0kb". Could you please help me with this problem

[calcervero]

I got this error, i don't understand it :c, can you help me?


Traceback (most recent call last):
File "GenData.py", line 102, in <module>
main()
File "GenData.py", line 41, in main
cv2.CHAIN_APPROX_SIMPLE) # compress horizontal, vertical, and diagonal segments and leave only their end points
ValueError: not enough values to unpack (expected 3, got 2)

[mikenguyen]

Thanks Chris, very good video and clear training for new guide as me

[cyilmusic]

is it possible to use this idea and train the program to recognize handwriting ?

[hanenbenneji]

Hello

I want to know what is the approach used to obtain basic learning from file 'GenData.py'. Is this a learning with neural networks?

I am waiting for your answer
thank you in advance
cordially

[abhisheknayak]

I'm getting a wrong output. The characters 'ABCD123' are recognised as Any help?

[fransdebruijn]

dude, get yourself a decent microphone, your send a shiver down my spine !

[danialahmed]

When I am trying to train it on my dataset I am getting empty array thus not writing any thing to classifcations and flattened image.txt?

[MarshalAmin]

Hey man, Thanks for the tutorial,
I Generate the Dataset correctly but when I'm running the Train and Test code, it produces only integer values, for instance, 486 instead of ABC, How can I fix it?
Thank you again

[eliyatech]

hey man this code have error in line 65 main file

[sibghatullah]

when we use multiple sentence in testing phase, the letters in output window are written in wrong sequence, can you help how to handle this issue?

[raulbastosferreira]

Hello, I'm Having this error on GenData:

cv2.CHAIN_APPROX_SIMPLE) # compress horizontal, vertical, and diagonal segments and leave only their end points
ValueError: need more than 2 values to unpack

How can I solve this?

And on TrainTest:


kNearest = cv2.ml.KNearest_create() # instantiate KNN object
AttributeError: 'module' object has no attribute 'ml'

[Madhu]

Hi Chris Dahms,

Great Video for New bees like me.... Could you please let me know how to train with more Fonts and Special chracters

[DivyaSingh-srqh]

How can I train it using a bigger data-set instead of training characters using an image? If it is possible, which data-set can be used?

[HansPeter-tshb]

Hi, Great videos! I did more GenData training after your file and it overwrote that previous data, so I copied that new data and added it to the bottom of the new data but then the Training program couldn't read the flatten image file. How do you create a larger training file?