Traditionally, this kind of technology requires very powerful graphic processors and is used only in the cloud. AIris Vision overcame this problem which became their main competitive advantage.

This AI evolutionary visual recognition startup from Romania created a technology that can recognize and understand objects and concepts in photos, videos and live streams. Their technology can run directly on low-end devices like mobile phones, raspberry pi, cctv camera, drones, etc.

They provide a software development kit (SDK) that can be used to quickly build an image-recognition-ready app or be integrated into existing apps, based on the specific classes that developers will learn with free API tools.

How does it work? For example, if someone wants to learn to detect fire, they have to upload images and videos with fire. Then they download the SDK that contains this class and use it for detecting fire in images, videos, live streams. – says Bogdan Grumezescu, co-founder of AIris

AIris is already in use for adult content moderation (automatically detect adult content) and identifying classes related with gas and petrol like pipeline maps, platforms tankers, drill pattern, well spacing maps.

Their main competition is Imagga and Clarifai, but AIris claims to have better accuracy and no GPU acceleration and doesn’t operate in cloud. They can recognize and classify objects in real time, using only one core of the CPU processor and also locate and track this objects.

AIris clients pay a one-time price per device, or/and a monthly fee per sever (SaaS , PaaS platforms) for image recognition in an unlimited number of frames from videos and photos. The market size for image recognition is now over 40 billion dollar. The team is currently self-funding the project but also actively looking for investors.