This is a python based optical character recognition tool (OCR) with GUI. The GUI is supported with python tkinter. The OCR part is handled by using py-tesseract which is a a wrapper for Google’s Tesseract-OCR Engine. This tool also provide a manual correction (provided by python-datamuse) using homophones of target words.
- Import one image from your computer at a time and OCR the imported image
- Select a part of the image and do OCR only on that region
- Can edit/copy/paste the result
- Provide homographs and homophones as other possible words
Once you've downloaded all files. Install following dependencies: (Tesseract, Pillow, Numpy, py-tesseract, python-datamuse)
An example using Homebrew, Anaconda, Pip to install aboving dependencies contains following command lines:
- brew install tesseract
- conda install numpy pillow
- pip install pytesseract python-datamuse
In your console, at directory /PyOCR/pyocr/ run below commend line:
- ./PyOCR.py
Or
- python3 PyOCR.py
to launch the GUI. Then you can select the image to OCR with. And you can also get a list of homographs and homophones of a word as its candidates.
Once you launch the GUI, you will see an interface like this:
Click the Browse then you can select the image you want to do OCR with. For example, below there, a scan of a business card is imported. And the OCR results are shown in the middle.
If you only need the information about this Dan Porter, you can drag a rectangular box on this image and click Update OCR, and the tool will only do OCR on text inside that box.
Here is an example showing the result with a photo and the fuzzy search on a word fast.
Another example with a screenshot of SMS.
