With the benefits of video processing technologies such as deep learning and artificial intelligence of increasing importance to the security industry, ensuring that infrastructure can cope with the processing load is critical. As a result BrainChip has launched a PCI-Express accelerator card to aid object recognition.
What is it?
BrainChip, a developer of software and hardware accelerated solutions for advanced artificial intelligence and machine learning, has released a hardware acceleration board, BrainChip Accelerator.
This is an eight-lane PCI-Express add-in card that increases the BrainChip Studio software’s speed and accuracy when carrying out object recognition functions by up to six times. The card also increases the support for simultaneous video channels in a system to 16 per card.
The accelerator card is very low-power and can be easily installed within existing video surveillance systems without upgrading power systems or thermal management requirements.
What does it do?
BrainChip Studio helps security-focused businesses and organisations to rapidly identify objects in large amounts of archived or live streaming video. By processing multiple video streams simultaneously, the BrainChip Accelerator add-in card enables those organisations using the software to search increasing amounts of video faster, with a higher probability of object recognition and lower total cost of ownership.
The system learns from single low-resolution images, which can be as small as 20 x 20 pixels, and performance with regard to recognition is claimed to be very accurate, even in low-light, low-resolution or noisy environments.
BrainChip states that the accelerator card is a milestone in the development of neuromorphic computing, a branch of artificial intelligence that simulates neuron functions.
The processing is done by six cores in a field-programmable gate array (FPGA). Each core performs fast, user-defined image scaling, spike generation and spiking neural network comparison to recognise objects. Each core can process up to 100 frames per second.
How fast is fast?
In comparison to GPU-accelerated deep learning neural networks, this approach represents a 7x improvement. BrainChip states that there is an estimated four exabytes of data stored in surveillance systems. Speed and accuracy of analysis are critical for security agencies, and the ability of BrainChip Accelerator to process video six times faster, while improving the accuracy of object recognition, is significant.