search

NJACKWinterOfCode / Printed-Text-recognition-and-conversion

A software for extracting text from scanned images of printed text documents
MIT License
4 stars 12 forks source link

Upgrade to PyQt5 #35

Open dom-k opened 5 years ago

dom-k commented 5 years ago

I tried to set up the project with PyQt4 but couldn't install PyQt4 with pip:

ERROR: Could not find a version that satisfies the requirement PyQt4 (from versions: none)
ERROR: No matching distribution found for PyQt4

Even after installing python3-pyqt4!

roshniRam commented 5 years ago

You will have to first download it's wheel then install it using pip. Download its wheel from here: https://www.lfd.uci.edu/~gohlke/pythonlibs/#pyqt4 Then go to the download location: pip install whl_file_name.whl

dom-k commented 5 years ago

Okay, I am developing on Linux, so these wheel files wouldn't work. 🙁 It would be more convenient to install dependencies from pip. How about upgrading to PyQt5? I've made the GUI version work with PyQt5, but when I run python main.py I get the following error:

Traceback (most recent call last):
  File "main.py", line 5, in <module>
    perform_ocr(".\image_samples\Test_.jpg")
  File "/home/dominik/Workspace/Printed-Text-recognition-and-conversion/ocr.py", line 27, in perform_ocr
    all_words, words_on_line, max_height_on_line = get_words(raw_image)
  File "/home/dominik/Workspace/Printed-Text-recognition-and-conversion/segmentation_words.py", line 24, in get_words
    img_for_det = image_for_detection(raw_image)
  File "/home/dominik/Workspace/Printed-Text-recognition-and-conversion/before_segmentation.py", line 9, in image_for_detection
    sm_image = cv2.GaussianBlur(raw_image,(5,5),0)
cv2.error: OpenCV(4.1.0) /io/opencv/modules/core/src/matrix.cpp:757: error: (-215:Assertion failed) dims <= 2 && step[0] > 0 in function 'locateROI'

Any ideas? 😕

roshniRam commented 5 years ago

Looks like it's some image size problem. I'll look into it.

dom-k commented 5 years ago

Okay cool, I'll create a PR. 👍