Programming Intelligence That Goes Beyond Pre-Built Models

Why I Chose C++/Python and Computer Vision

For the simplest machine learning tasks, such as vehicle recognition/detection or segmentation, today you don't need to be a programmer, everything has already been done – YOLO as a sample. Teaching a machine something new - is a whole other level (i don't mean COCO's database).

For successful learning computer of vision, requires a deep understanding of how a neural network obtains the knowledge to recognize a specific object ...dimensions, angles, lighting, and so on. It is necessary to catch not only what helps learning, but also what hinders it.

I am fascinated about constructing of logic, it is much more exciting than playing chess. It is a challenge that gives meaning to life. It's so exciting that sometimes you forget to sleep ...but you know that you definitely need to have some sleep or take a walk when the construction of logic stalls. The greatest pleasure is end result, you know you did it.

Programming language - aren't the same as logic. Today, a neural network can write a code for you. For the code to get working, logic is needed; the coding needs to be combined into one chain, integrated in a certain sequence, when united different hardware performs one function, and shares information to each other.

My first project, build from scratch, was really challenging. But I was so captivated by it that it only took me three weeks to finish, using only C++ language.

Today i have united two of mains future of automations - logic and digital intelligence ...completely different possibilities.

Our digital department: https://www.bodycaststudio.co.uk/digital-automations

Thomas J.