Image processing is now a hot topic. It helps us to process images, analyse them, even to manipulate them and generate data from it. There's a branch of image processing that I just learned yesterday and that is OCR or Optical Character Recognition.

OCR is widely used today, for example to digitize old books that already lost their copyright. Those books are now free and open-source (e.g. Project Gutenberg)

My idea is to make a personalised OCR for my handwritten notes. Instead of the fonts from books being recognised, why not our own handwritings?


Q: Why don't you just simply type on your laptop at classes?

A: First, I don't have a laptop. I only have a computer at home. Second, I read that writing is good for our brain compared to typing. So I don't want to lose this good habit. It can make us remember the subject better.

After I read some tutorials, I can summarise what I will do.

1. I will have 62 folders (26 lower case alphabet + 26 upper case alphabet + 10 digits number)

2. Each folder will contain samples of my handwriting in variations.

3. Do a simple KNN training.

4. Test the model and evaluate.