The DeepEyes Algorithm
Born from almost two decades of research conducted by video analytics and neuro technology scientists as well as experts in predictive technologies we have developed a complex, self-learning algorithm that can reliably and accurately analyze video data in real-time without external communication like grid computing or neural networks. Instead the technology is an on-device neural network accellerator.
The technology operates on standard off-the-shelve hardware, even on a tiny micro-controller.
The uses of our technology are endless and we continue to push the limits of what is possible.
A unique combination of mathematical and statistical algorithms
Processing of information in RAM
Use of low-level languages
What makes DeepEyes unique
Processing of large amounts of data per unit cycle.
We solve complex classification tasks on low-end controllers (e.g. Arduino)
Currently testing solutions that allow us to increase our speed 20 times. For example, 4000 polygons are classified in a few milliseconds.
Most analytics tasks around our technology can be solved using a microcontroller
Less Code – More Maths
The DeepEyes technology uses several combinations of Bayesian methods for the classification and analysis of images and data. The technology is adaptive, this means it can evolve models that dynamically respond to changes. For example, the core algorithm for our emotion recognition technology consists of definitions of about 98 physiological parameters. It is able to work without external communication; that means that it can operate on-device and does not need neural networks or grid computing. Depending on our customers´specific requirements all our algorithms are slightly altered or combined with additional technologies.
DeepEyes includes, but is not limited to
Patterns of behavior
Reactions to events
Moving objects of any kind
Counting of objects and persons
Anomalies and outliers recognition
Pulse, eye movement, respiratory rate recognition
Gesture and Motion Recognition
The DeepEyes gesture and motion recognition accurately and completely tracks precise body movements through a normal HD camera. Depending on customer specifics the technology focusses on specific body parts and reaches highest accuracy in identifying defined outliers, patterns or anomalies.
The generic core model accompanies the human body down to hand and toes and recognizes all kinds of body movements down to single finger movements. More generic models have been developed based on the core model and will be customized to customer requirements.
All this is possible with a standard web-camera on standard hardware. No additional equipment or accessories are required. The monitored individual moves freely without wiring which makes traditional motion sensors obsolete. Also, the Deepeyes Gesture and Motion Recognition allows for endless sophisticated tracing and tracking of required process steps in all manufacturing and logistics requirements.
Up to eight different recognitions can be assembled on one solution. Object and gesture/motion recognitions can be assembled in one solution.
only selected pixels are being processed
data are not being stored
data are not sent via the cloud
faces can be blurred
biometrical data are not relevant and are not being analyzed
Object Recognition and Monitoring
The DeepEyes core object recognition solutions identify even minor changes, anomalies or outliers on surfaces from raw materials and purchased parts to partially completed goods to finished goods or merchandise.
Also, being an artificial intelligence technology in the true sense of the word DeepEyes learns and recognizes any kind of flaws even in irregular shaped objects, like produce or meat.
DeepEyes is ideal for replacing costly manual labour or maintenance prone high-tech sensors and machinery. Up to eight different recognitions can be assembled on one solution. Object and gesture/motion recognitions can be assembled in one solution.