husseinyoussef / arabic-ocr Goto Github PK

OCR system for Arabic language that converts images of typed text to machine-encoded text.

License: MIT License

Python 100.00%

ocr character-segmentation arabic machine-learning segmentation image-processing dataset neural-network opencv-python scikit-learn

arabic-ocr's Introduction

Arabic OCR

OCR system for Arabic language that converts images of typed text to machine-encoded text.
The system aims to solve a simpler problem of OCR with images that contain only Arabic characters (check the dataset link below to see a sample of the images).

Important Note

The system currently supports only letters (29 letters) ا-ى , لا (no numbers or special symbols).

Setup

Install python then run this command:

pip install -r requirements.txt

Run

Put the images in src/test directory
Go to src directory and run the following command
```
python OCR.py
```
Output folder will be created with:
- text folder which has text files corresponding to the images.
- running_time file which has the time taken to process each image.

Pipeline

Dataset

Link to dataset of images and the corresponding text: here.
We used 1000 images to generate character dataset that we used for training.

Examples

Line Segmentation

Word Segmentation

Character Segmentation

Testing

NOTE: Make sure you have a folder with the truth output with same file names to compare it with the predicted text.

From within src folder run:

python edit.py 'output/text' 'truth'

Performance

Average accuracy: 95%.
Average time per image: 16 seconds.

NOTE

We achieved these results when we used only the flatten image as feature.

References

arabic-ocr's People

Contributors

Stargazers

Watchers

arabic-ocr's Issues

TypeError: Cannot cast array data from dtype('uint32') to dtype('int32') according to the rule 'safe'

Hey when i m loading one of your test images i m getting this error any idea how to solve this ?

Can you solve error (cv2.error: Unknown C++ exception from OpenCV code) ?

C:\Users\AMR\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.9_qbz5n2kfra8p0\LocalCache\local-packages\Python39\site-packages\sklearn\base.py:329: UserWarning: Trying to unpickle estimator MLPClassifier from version 0.22 when using version 1.0.2. This might lead to breaking code or invalid results. Use at your own risk. For more info please refer to:
https://scikit-learn.org/stable/modules/model_persistence.html#security-maintainability-limitations
  warnings.warn(
  0%|          | 0/3 [00:02<?, ?it/s]
multiprocessing.pool.RemoteTraceback: 
"""
Traceback (most recent call last):
  File "C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.9_3.9.2800.0_x64__qbz5n2kfra8p0\lib\multiprocessing\pool.py", line 125, in worker
    result = (True, func(*args, **kwds))
  File "C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.9_3.9.2800.0_x64__qbz5n2kfra8p0\lib\multiprocessing\pool.py", line 48, in mapstar
    return list(map(*args))
  File "C:\Users\AMR\Scrapy_projects\Arabic-OCR\Arabic-OCR\src\OCR.py", line 23, in run2
    char_imgs = segment(line, word)
  File "C:\Users\AMR\Scrapy_projects\Arabic-OCR\Arabic-OCR\src\character_segmentation.py", line 659, in segment
    valid = filter_regions(binary_word, no_dots_copy, SRL, VP, upper_base, lower_base, MTI, MFV, top_line)
  File "C:\Users\AMR\Scrapy_projects\Arabic-OCR\Arabic-OCR\src\character_segmentation.py", line 401, in filter_regions
    cc, l = cv.connectedComponents(1-(no_dots_copy[:, end_idx:start_idx+1]), connectivity=4)
cv2.error: Unknown C++ exception from OpenCV code
"""

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "C:\Users\AMR\Scrapy_projects\Arabic-OCR\Arabic-OCR\src\OCR.py", line 98, in <module>
    running_time.append(run(images_path))
  File "C:\Users\AMR\Scrapy_projects\Arabic-OCR\Arabic-OCR\src\OCR.py", line 47, in run
    predicted_words = pool.map(run2, words)
  File "C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.9_3.9.2800.0_x64__qbz5n2kfra8p0\lib\multiprocessing\pool.py", line 364, in map
    return self._map_async(func, iterable, mapstar, chunksize).get()
  File "C:\Program Files\WindowsApps\PythonSoftwareFoundation.Python.3.9_3.9.2800.0_x64__qbz5n2kfra8p0\lib\multiprocessing\pool.py", line 771, in get
    raise self._value
cv2.error: Unknown C++ exception from OpenCV code

detect number in image

hi
is your code support detect number in this picture ?

OCR unable to read the input image for text detection uint error

I have tried to detect Arabic from Aljazeera news snaps but the code failed to read the file. The code is restricted to read 8- bit depth jpeg or png image. However, I have used the 24-bit depth jpeg image and 8-bit bitmap image (added .bmp in list) but still that uint error. Is your code is supporting a specific Arabic font? and how to resolve this bit depth issue. Please guide thanks.

ValueError: too many values to unpack (expected 2)

Operating System: Windows 10 - 64 bit
Python 3.7
after installing requirements.txt
when I run python OCR.py I get the error as mentioned in the subject: ValueError: too many values to unpack (expected 2)

train model

how do i train model?

Adding docker

Hello,
Thanks for the great tool, I am trying to build this on an M1 Macbook but getting errors on installation
error: the clang compiler does not support 'faltivec', please use -maltivec and include altivec.h explicitly.

I tried to add docker file but still got the same issue.

FROM python:3.8-slim-buster
WORKDIR /app
COPY requirements.txt requirements.txt
RUN pip3 install -r requirements.txt
COPY . .

CMD [ "python" ,"OCR.py"]

Handwritten support?

I just looked at the datasets that you used and found that it only use computer typed text not handwritten by a human, is this correct?!

Train the model

Hey Hussein, you have made a good project but i want to ask one thing, how to train the model for arabic image/handwritten image or how you trained the model?
Answers are appreciated :)

Thank you.

Training Process is stuck after breakpoints

After reading all images from dataset the training process is stuck on scores[ ] Pdb. Please describe is the training process is running in background? or debugger is unable to continue the process. Moreover, how much time will it require to train the models?

error when running ocr.py

please help me to reslove this problem.thank you!

ValueError: With n_samples=0, test_size=None and train_size=0.8, the resulting train set will be empty.

Hello,

I try to run the training but I get this error in below:

File "train.py", line 214, in <module>
    train()

  File "train.py", line 143, in train
    X_train, X_test, Y_train, Y_test = train_test_split(X, Y, train_size=0.8)

  File "C:\ProgramData\Anaconda3\lib\site-packages\sklearn\model_selection\_split.py", line 2120, in train_test_split
    default_test_size=0.25)

  File "C:\ProgramData\Anaconda3\lib\site-packages\sklearn\model_selection\_split.py", line 1805, in _validate_shuffle_split
    train_size)

ValueError: With n_samples=0, test_size=None and train_size=0.8, the resulting train set will be empty. Adjust any of the aforementioned parameters.

Please advise!
Thank you in advance.

need working example

i'm trying to use this code, but i get empty output. I think there is something wrong with my image,not sure. can you provide some sample working test image

Questions

Hey man
I want to ask you a few questions
Do you have a skype account or anything like that where i can message you
I willl not take a lot of your time
Thanks in advance

add support to to dectect arabic numbers (Eastern Arabic numerals)

install error

hi
in install it say :

pip3 install -r requirements.txt
Defaulting to user installation because normal site-packages is not writeable
Requirement already satisfied: numpy in /home/mohsen/.local/lib/python3.6/site-packages (from -r requirements.txt (line 1)) (1.18.3)
Requirement already satisfied: opencv-python in /home/mohsen/.local/lib/python3.6/site-packages (from -r requirements.txt (line 2)) (4.2.0.34)
Requirement already satisfied: scikit-learn in /usr/lib64/python3.6/site-packages (from -r requirements.txt (line 3)) (0.22.2.post1)
Requirement already satisfied: scikit-image in /home/mohsen/.local/lib/python3.6/site-packages (from -r requirements.txt (line 4)) (0.16.2)
Requirement already satisfied: pandas in /home/mohsen/.local/lib/python3.6/site-packages (from -r requirements.txt (line 5)) (0.25.3)
Requirement already satisfied: matplotlib in /home/mohsen/.local/lib/python3.6/site-packages (from -r requirements.txt (line 6)) (3.2.1)
Collecting multiprocessing
  Using cached multiprocessing-2.6.2.1.tar.gz (108 kB)
    ERROR: Command errored out with exit status 1:
     command: /usr/bin/python3.6 -c 'import sys, setuptools, tokenize; sys.argv[0] = '"'"'/tmp/pip-install-oe37fbft/multiprocessing/setup.py'"'"'; __file__='"'"'/tmp/pip-install-oe37fbft/multiprocessing/setup.py'"'"';f=getattr(tokenize, '"'"'open'"'"', open)(__file__);code=f.read().replace('"'"'\r\n'"'"', '"'"'\n'"'"');f.close();exec(compile(code, __file__, '"'"'exec'"'"'))' egg_info --egg-base /tmp/pip-install-oe37fbft/multiprocessing/pip-egg-info
         cwd: /tmp/pip-install-oe37fbft/multiprocessing/
    Complete output (6 lines):
    Traceback (most recent call last):
      File "<string>", line 1, in <module>
      File "/tmp/pip-install-oe37fbft/multiprocessing/setup.py", line 94
        print 'Macros:'
                      ^
    SyntaxError: Missing parentheses in call to 'print'. Did you mean print('Macros:')?
    ----------------------------------------
ERROR: Command errored out with exit status 1: python setup.py egg_info Check the logs for full command output.

Missing text files in 'truth' Folder when Running ocr.py on handwritten images

Challenge
I am experiencing an issue with the ocr.py file in my cloned repository. After following the provided instructions to have the images I want to convert to text in the test file, I ran the ocr.py file, but I'm not getting any output in the "truth” folder or anywhere else as expected.

Steps to Reproduce

Clone the repository.

Follow the instructions to prepare the images in the test folder.

Run the ocr.py file using python path/ocr.py

Observe that no output is generated in the "truth" folder

Wrong detection

ouput:
احسد ممبطش محهد حدعشيشي كغا لي معدطفذس ح ن خثعي