The University of South Carolina is offering licensing opportunities for Facial Recognition Based on Converted Spiking Neural Network
The current method for facial recognition requires a lot of computing resources (high-performance computers) which is not portable, and the accuracy is very limited.
We developed computer software with an innovative method to recognize the same person from different pictures. So, we can identify who is who based on the photos or videos from different sources. The model we developed is a very efficient model that does not require a lot of computing resources and can be loaded into mobile devices, such as cellphones, tablets, computing sticks, and mobile computing boards to make it portable. The accuracy of our model is equal to or even better than those existing model that requires a high-performance server or high-end GPU/CPU.
The potential market will be the first responders for security screening at airports, ports, railroad/bus stations, shopping malls, etc., and similar highly populated areas.
Advantages and Benefits:
Our production requires significantly less computing power and battery power for achieving better recognition accuracy. The model we developed can be used to deploy to mobile devices for high portability.