Brainchip break through...

Brainchip break through published:

https://arxiv.org/pdf/2404.08858

“A Lightweight Spatiotemporal Network for Online Eye Tracking with Event Camera

Yan Ru Pei Sasskia Bru ̈ers Sebastien Crouzet Douglas McLelland Olivier Coenen

{ypei, sbruers, scrouzet, dmclelland, ocoenen}@brainchip.com Brainchip Inc.23041 Avenida De La Carlota, Suite 250, Laguna Hills, CA 92653

Abstract
Event-based data are commonly encountered in edge computing environments where efficiency and low latency are critical. To interface with such data and leverage their rich temporal features, we propose a causal spatiotempo- ral convolutional network. This solution targets efficient implementation on edge-appropriate hardware with limited resources in three ways: 1) deliberately targets a simple architecture and set of operations (convolutions, ReLU ac- tivations) 2) can be configured to perform online inference efficiently via buffering of layer outputs 3) can achieve more than 90% activation sparsity through regularization duringtraining, enabling very significant efficiency gains on event- based processors. In addition, we propose a general affine augmentation strategy acting directly on the events, which alleviates the problem of dataset scarcity for event-based systems. We apply our model on the AIS 2024 event-based eye tracking challenge, reaching a score of 0.9916 p10 ac- curacy on the Kaggle private testset”

Brainchip advancing the practicality of using event based cameras and becoming more essential to Prophesee and companies like Samsung, Sony, Qualcom as well as Aerospace and Defence.

My opinion only DYOR

Fact Finder

PS: For those who remember a retired engineer who posted as Diogenese here you might remember also that a poster Pigs Might Fly rubbished him up hill and down dale when he suggested that Brainchip’s research would likely lead to this break through. He who laughs last etc;