top of page

Ville Karlsson

Spiking neural networks for vision and image processing

Doctoral Student

Spiking neural networks (SNNs) are a step closer to biological neural networks. They work in a fundamentally different way than classical neural networks and are so titled the 3rd generation of neural networks.

Recent advances on the hardware side have successfully demonstrated advantages of SNNs. When run on a neuromorphic processor, a SNN can attain 100x higher energy efficiency and significantly lower latencies when compared to a classical neural network approach. This makes them a great candidate when working with on-device NN models that often are reliant on battery and may need to quickly respond to changes in the environment.

In this project, the aim is to find high quality, high efficiency and low latency SNN based solutions to computer vision problems.

Academic supervisor
Joni Kämäräinen
Industry partner
Samu Koskinen

Huawei

Spiking neural networks (SNNs) are a step closer to biological neural networks. They work in a fundamentally different way than classical neural networks and are so titled the 3rd generation of neural networks.

Recent advances on the hardware side have successfully demonstrated advantages of SNNs. When run on a neuromorphic processor, a SNN can attain 100x higher energy efficiency and significantly lower latencies when compared to a classical neural network approach. This makes them a great candidate when working with on-device NN models that often are reliant on battery and may need to quickly respond to changes in the environment.

In this project, the aim is to find high quality, high efficiency and low latency SNN based solutions to computer vision problems.

bottom of page