Coding Challenge

How to run the program

• Install the python package opencv-python in your system/ python virtual environment

• Go to the directory and run the programs by executing the following commands. Replace python3 with the python command installed in your system. Both are independent of each other and are meant to be run seperately.

    python3 code.py
    python3 code2.py
  

• Close any windows that pop up and continue the execution by pressing any key on the keyboard.

Two implementations

• code .py

  • Implementation based on simple parsing.
  • Accuracy better than the other implementation but still not perfect.
  • Results acheived using very narrow boundary conditions heavily dependent on particular pixel value thresholds. We might want to tweak these values/condiitons for other inputs to maximise accuracy.
  • Results saved in letters directory

• code2 .py

  • Initial approach.
  • Implementation based on techniques of thresholding and segmentation.
  • Not many variables involved in boundary conditions since the parsing is done in a thresholded image (because there are only two distinct pixel values present in the image after thresholding, 0 and 255).
  • Less accurate results.
  • Results saved in letters2 directory