Skip to content

printed text to speech conversion by improving OCR accuracy using spell correction

Notifications You must be signed in to change notification settings

harshablaze/Printed-text-to-speech-using-ocr-and-spell-correction

Repository files navigation

NOTE: I developed this project on Windows 10 X64bit OS and Python 3.9.2 X64bit version python Environment
if you want to run this project on linux change the paths and tesseract path in ocr.py


Setting up the project

1. After Downloading the project

  • src>cd 4-4-project

2. activate the env

  • src\4-4-project>.\env\Scripts\activate

NOTE: if your env is activated you get (env) before your src path if env failed to activate then follow step 3 else go to step 4

3. install requirements.txt

  • src\4-4-project>pip install -r requirements.txt

Run the project

4. to start the project

  • src\4-4-project>python start.py


after start is running acquireimage will run automatically
press spacebar to capture image else Esc to enter image number manually
for manual input first save image in images folder with sample45.jpg name (any number instead of 45 is ok)
now if you pressed Escape button while acquiring image it asks for image number so enter 45 or the number you saved
you can modify code to take manual input images from your custom folder

while execution some images may popup to show internal process press ctrl+W or Esc to proceed


Troubleshooting errors

1. if you encountered errors in installing requirements.txt

  • check python architecture
src\4-4-project>python
>>import platform
>>platform.architecture()

output: ('64bit', 'WindowsPE')

2. if errors encountered after executing start.py

  • check if it is package error or input files error or path errors and rectify manually

any issues connect with me through linkedin

Releases

No releases published

Packages

No packages published

Languages