We built a model that finds anomalous areas on the belt. Our system detects little scratches and other defects in real-time.
We’ve made a system for detecting wall cracks based on drone images and estimating the combined length of these cracks. We accurately detect cracks starting from a 0.1 mm width.
We’ve created a drone that is capable of reconstructing 3D scenes from an onboard camera. We are currently working on autonomous navigation system.
We've detected intrinsic periodicity of the data, correlations between parameters and rare anomalous behavior of the 4G network for one of the mobile industry leaders.
Anomaly detection in mining machines for preventing costly damage during operation. The anomaly detector is based on the onboard computer data from various sensors. The system detects deviations and notifies operators.
Given a continuous stream of device logs, the system determines how strange is current device behavior as it can be a flag of soon breakdown. The system works in runtime with logs from more than 0.5M devices. So we can predict device faults.
