Recognition
system
Recognition systems are used to automate visual inspection processes. Our neural network analyses the photo or video stream and makes conclusions according to the specified criteria.

The algorithm can help in tracking the movements of people and vehicles, checking the quality of finished products and documents, controlling production processes and equipment, recognising faces and numbers and many other tasks.

The program works remotely or at the plant, also we have experience in monitoring space using drones. The results of the analysis are transmitted to clients via the api in any convenient format and can be integrated as part of the management software.

- Recognition system

Recognition system
Artificial intelligence system developed by True Detection
  • Recognition and analysis of objects

  • Data analysis and real-time report generation

  • Neural Network Analytics – self-developed software

  • Python is used for programming

The uniqueness of the architecture
  • Customised neural network architecture
  • Based on CenterNet with custom backbone
  • Multiple layers
  • Convolutional Neural Network (CNN)
  • Deep learning
  • Advanced feature extraction techniques
  • Advanced approaches to architecture: FPN, Separable Convolution, Deformable Convolution, СE-block, etc.
Uniqueness of learning
  • Own data collection methodology
  • Dataset from real sorting plants
  • Automatic regular updates of the image database
  • Continuous learning

OPERATION CONCEPT OF THE RECOGNITION SYSTEM

Advantages of the system
No. 1
Recognition accuracy - 80%
No. 2
High speed of operation
No. 4
Works with any type of connection (cameras on the factory, on the street, etc.)
No. 5
Universal output data format (api access, application integration, management software)
No. 3
Continuous registration of indices

- Reporting system

Reporting system
Advantages:
  • Real-time reporting
  • User-friendly interface
  • Convenient filters and data sorting
Opportunities:
  • Tracking company performance indicators
  • Data in graphs in the interface
  • Uploading detailed reports in Excel format
  • Individual configuration of the reporting system
UAE, Abu Dhabi, Masdar City, Incubator Building
Made on
Tilda