As developers rush to build AI based systems, users can discover the incredibly high cost of some of these systems – often only after they are deeply committed. The real cost is not necessarily in the price of the software which may sometimes be “given away for free”, but in the cost of the associated infrastructure.

As an early pioneer in the field of Autonomous Multi-sensory AI, iOmniscient has designed its systems to minimize infrastructure costs to the point where an intelligent system can often cost less than a simple recording system.

This is achieved at three levels:

• • Being able to process inputs from fewer, less expensive sensors. Up to 50X fewer IoT sensors may be required
• • Through using less computing infrastructure. Systems can be designed to require 90% less storage and network bandwidth, and no GPUs. The ability to do AI without GPUs reduces both costs and power requirements which can be reduced by 16X making the entire system sustainable over the longer term.
• • By automating manual processes to minimize human involvement. Autonomous systems can require up to 15X less manpower

Savings can pay for the entire intelligent system

Using less expensive sensors

The system is designed to provide accurate analytics even when the inputs are of poor quality. It can detect objects even when they are invisible to the human eye. (See: video example – detecting an abandoned bag on a 0.025 megapixel camera).

It can anonymously track individuals at a distance of 20 to 30 meters in a crowd on very low resolution cameras, understanding their demographics – using just a 1 or 2 megapixel camera. This can be extremely valuable to retailers to understand their customers without infringing their privacy.

And now with the its recent patented development involving analytics ON THE MOVE where some analytics can be performed on drones and body worn cameras, the overall number of cameras required is greatly reduced often by more than 20 times what would be ordinarily required.

Less Storage, less Network Bandwidth and no GPUs.

iOmniscient’s Smart Compression capability reduces storage and network bandwidth requirements by around 90% on top of what can be achieved with the industry’s most effective standard codec, namely H.265. (See: video example). Further iOmniscient systems can operate without GPUs greatly reducing the over cost of the computing as well as power consumption. For a large system the power requirement and hence the carbon footprint of the system can be reduced by a factor of 16 making the entire operation much more sustainable.

Minimizing Human Involvement

The AUTONOMOUS nature of iOmniscient systems where the camera itself can locate the nearest appropriate responder and tell them where to go and what to do, enables users of such systems to minimize the number of people they require to man control rooms. For large sites, they would generally require 95% fewer people. Even for small sites they would normally be able to reduce their manpower requirements by 60 to 70%.

To operate a system autonomously the system needs to be extremely accurate and this is achieved through the NAMS module (Nuisance Alarm Minimization System) that arms every application.

The savings of 15 times or more in terms of reduced manpower due to the autonomous nature of the system is rarely considered when the system is purchased but is a major saving in cost over the life of the system.

Proven Savings

These infrastructure saving designs are built into every one of iOmniscient’s 300+ AI Use Cases that operate across 30+ industries. Hundreds of projects and 75 Smart Cities have benefited from them already.

Savings in Infrastructure pays for the system

The savings in cameras, storage, computing resources and manpower are such that they essentially make the system less expensive than a recording system and significantly cheaper than any other intelligent system on the market and the ability to operate with GPUs not only reduces the cost but makes the system sustainable.

This ability to reduce infrastructure costs for AI systems is protected by a large number of international patents and is therefore not available from anyone else.