Video-Based Artificial Intelligence 2020-07-10T09:49:04+01:00

Video-based A.I. recognition for real-time process monitoring

A new level of efficiency and safety on the shop floor

The Challenge

According to the Six Sigma method, the industry average for error costs amounts to 15-20 percent of sales. When companies have more than 30 percent error costs, they are often no longer competitive.

  • The smallest (avoidable) errors in logistics and production can sometimes cause high follow-up costs.

  • In many cases, errors are not detected early or effectively enough.

  • As a rule of thumb, the earlier an error is detected, the lower the follow-up costs.

The Solution

DeepEyes AI video-analytics technology detects errors in real time. These can be product, process and behaviour errors.

DeepEyes detects:

  • Safety or production risks arising from unusual or unwanted movements or gestures

  • Scratches, deformations, faulty packaging, faulty assembly, number of objects, filling levels, anomalies in shape and color

  • Anomalies in irregularly shaped items like fruits, vegetables meat that would otherwise need to be inspected by humans

Real-time alerts enable quick error corrections. This means that production delays, downtimes, recalls, recourse claims and potential reputational damage can be largely avoided.

Gesture and motion recognition helps to prevent accidents at work.

  • Efficient – DeepEyes captures up to eight “recognitions” in one solution, depending on specific customer needs.

  • Innovative – DeepEyes sees behaviour risks thus contributing to more safety and efficiency in the workplace.

  • Adaptable – The generic DeepEyes technology is adapted to capture very specific recognition needs.

  • Cost effective – DeepEyes runs on standard hardware and standard cameras. Maintenance is easy.

  • Safe – DeepEyes can work standalone and offline. This prevents data and privacy breaches.

  • Flexible – embedding, analyzing and enriching data is easy. Alerts can be sent to any device.

  • Hardware agnostic: DeepEyes solutions can run on existing cameras and hardware.

Based on our generic technology we build predictive modules for our customers´visual surveillance and monitoring requirements on the shop floor or anywhere, where worker safety as well as efficiency have highest priority.

Our videos give an idea of the accuracy of our models.

Data Security

DeepEyes solutions run on standard hardware and standard cameras. The algorithm operates in real-time and does not require cloud access for the analysis of video streams. Instead, it is edge-capable and can be used standalone/on-device.

Privacy Security

We do not collect biometric data nor do we store data. We guarantee data protection according to GdPR specifications.

Our technology does not require neural networks but exclusively relies on smart algorithms.

Object Recognition

DeepEyes detects and alerts even the tiniest flaws on the surface of moving objects

DeepEyes quality monitors objects of any kind.

DeepEyes works with the highest accuracy on off-the-shelf cameras and can even detect anomalies and deviations on moving objects, e.g. on a conveyor belt.
Read More

Motion Recognition

Based on customer requirements DeepEyes recognizes outliers in motions.

DeepEyes recognizes human motions of any kind

DeepEyes early indicates anomalies in human behaviour, e.g. deviations from standard procedures, individuals trespassing danger zones.
Read More

Gesture Recognition

DeepEyes accurately identifies exact hand movements, even down to every finger.

Recognize motions of hands and fingers

DeepEyes tracks precise body movements with standard hardware and HD cameras.
Read More


Over the years, based on customer requirements, we have further refined our technology.

Counting, Attire, Size

DeepEyes has various specific models which can serve as a basis for your unique recognition needs, e.g. counting of objects, assessing ability to perform duty. Please contact us.

If you would like to speak with our experts

Contact Us