This app uses Node.js and Python to accept image, get coordinates of bounding boxes (where text is likely present) and then draws bounding boxes on the top.
The initial screen should look like this:
click Choose File, then upload an image. Then click Submit.
Now you should see:
Click the hyperlink below. (Warning: this process would take quite a bit of time, depending on your image.)
After the wait, you should see:
Now the image is annotated with bounding boxes around the potential texts! Are they the same picture?
- The python script is also capable of generating texts as well. Add code to parse texts and display?
- Optimization
- Deployment - I tried to deploy this on Heroku, but was unsuccessful as Heroku instances are very low-memory and this process requires a lot of memory. Try to optimize?