To date, deep learning technology has been used to solve many narrowly focussed artificial vision problems such as character recognition or object detection, however experience has shown that biological visual systems are vastly superior in terms of speed and performance.
Neuromorphic cameras are a new visual sensing technology that offer the ability to ‘see’ in a manner like human eyes; these have been developed based on neuro-biological principles of retinal processing. Our novel neuromorphic processing algorithms are combined with deep learning technology to demonstrate significant improvements in real-time motion detection performance.
The use of neuromorphic cameras and our software solution overcomes many issues related to the processing of vast amounts of high-speed motion data whilst utilising limited embedded power (which is highly important for power restricted applications such as in-car computers).
As the demand for autonomous vehicles, driver assistance technology and associated systems continues to rise, manufacturers and providers are seeking to use advanced vision systems and artificial intelligence to help advance their products beyond current technologies.
NeuroEye has developed artificial intelligence algorithms for rapid motion detection using state of the art neuromorphic cameras, overcoming many of the difficulties with traditional vision technologies such as slow processing, fast motion, and distant moving objects. NeuroEye is at the forefront of an emerging high value market and will play a pivotal role in the adoption of neuromorphic sensors within automotive designs and manufacturing processes.