Visualizing the characters in an OCR database

We will now look at how to use neural networks to perform optical character recognition (OCR). This refers to the process of identifying handwritten characters in images. We have always been particularly sensitive to the problem of the automatic recognition of writing in order to achieve a simpler interaction between humans and machines. Especially in the last few years, this problem has been subject to interesting developments and more and more efficient solutions thanks to a very strong economic interest and an ever-greater capacity to process data of modern computers. In particular, some countries, such as Japan, and Asian countries in general, are investing heavily, in terms of research and financial resources, to make state-of-the-art OCR.

...