Security Surveillance

Problems facing large surveillance projects:

Need bulk decoding with AI.

Connect 100’s of cameras and use AI filtering to sort volume.

Upgrade to 4K resolution.

Legacy projects require higher resolution without replacing cameras.

Must cut operating costs by 80%.

Reduce power draw and reduce total size of storage archive.

Ultra dense decoding.

Manage 100's of camera feeds with AI.

A single Quadra VPU chip (T1U or T1A) can decode 48x 1080p30 live streams. Considering 1RU Quadra Video Server configuration may house up to 10 VPUs, our decoding capacity far out performs what a GPU is capable of for advanced AI analytics operations.

Quadra VPUs have a high-performance hardware decoder capable of decoding codecs: H.265 (HEVC), H.264 (AVC), VP9 and JPEG.

Upgrade system without replacing legacy cameras.

Phew. Large-scale legacy projects can keep their existing cameras because the AI enhancements run inside the VPU chip, processing your incoming video streams and adding image sharpening and facial recognition after the video is captured.
Ideal for any residential, commercial or industrial application.

AI inferencing

Key Frame Detection identifies alerts
so agents only focus on the priorities.

Increase your capacity by 10x simply by filtering out 90% of the noise. VPUs streamline video analysis and eliminate bottlenecks with high-efficiency prescreening security models. The key frame detection performs spot analysis on every 10th frame to identify key objects. These escalated alerts go to agent for closer review and intervention
Quadra demonstrates low-latency processing for real-time applications while scaling for workload demands.

4K & 8K high resolution

The decoder is designed to handle resolutions up to 8192×8192 pixels and supports smooth playback of 4K content at 125 frames per second (or 8K resolution at 60 fps when decoding 10-bit YUV video).
This capability ensures that Quadra VPUs can handle the most demanding video decoding tasks and high-quality applications.

80% Smaller archives

Archive high-quality surveillance footage without the bulk. 

Typically, the storage capacity of a video surveillance system is directly correlated to the CAPEX and OPEX investment; the more you spend, the more you get. That’s not the case here.
VPUs use AI video compression to reduce the bit rate of whole archives with zero quality loss. This enables nDVR storage to reduce the content to only 20% of its volume. Voila!

Reduce costs.

Lowest power consumption.

Hardware now beats software in consuming the least amount of energy. ASIC VPUs are engineered for 100% processing video. Tasking VPUs with ths single operation simplifies and accelerates the workflow by freeing up the CPU to do other video tasks only it can do, resulting in massive 10X more throughput at 1/10 th the cost.

In fact, incorporating VPUs in your video workflow will relieve 
8 of 10 servers, enabling them to be re-purposed or eliminated. 


80% less hardware uses 80% less energy.