Technology

The concept

 

Neural network technology and performing tasks of pattern recognition

 

The overall concept of pattern recognition is based on three main components: 

 

A sensor of perception (Camera, microphone, and other types suitable for application contexts: EEG, ECG, magnetic fields, …)

A mixed architecture, Von Neumann (or Harvard) and Neural for data acquisition, processing and sorting information, decision-making.

Inputs/Outputs in order to customize the board and to interact easily with external systems.

 

The future is now!

 

What is a neural processor ?

 

A neural chip is a processor that "imitates" the reasoning of the human brain. The main difference between the human brain and a neural processor is the limited number of "neurons" simulated by the processor.

 

The recognition engine, trained with the customer’s knowledge can recognize, with astonishing simplicity and in real-time, a pattern or a signal due to its parallel neural architecture.

 

 

 

A radical change of approach for calculation: from high performant CPU for many sequential access to memory, to a processor dedicated to each input memory (associative memory)