Building Computer Vision Projects with OpenCV 4 and C++

Preparing the data

We have 30 characters and numbers, distributed along 702 images in our dataset with the following distribution. We can check that there are more than  30 images for numbers, but some letters such as K, M, and P, have fewer images samples:

In the following image, we can see a small sample of images from our dataset:

This dataset is very small for deep learning. Deep learning requires a huge amount of samples and is a common technique. In some cases, use a dataset augmentation over the original dataset. Dataset augmentation ...