Overview

The Eyeflow.AI Video Analytics platform was developed by SiliconLife to enable the creation of market applications using Artificial Intelligence in a practical and agile way.

Several native features:

• Configurable and modular detection conditions
• Friendly interface:
  • Construction of operational flows with drag’n’drop of AI components
  • Complete development environment for new applications
  • Extensibility for adding / developing new components
• Agile management of new data requirements and business processes
• Simplified integration with other platforms
• Simple and consistent integration with legacy data systems
• Modular, orchestrated design for high call volumes
• Easy integration of open software components, new AI algorithms or with cloud services - Configuration of different types of alarms and reports - Integration with corporate Data Lake

EyeFlow offers simplified access to Artificial Intelligence solutions for video analytics in a simple way, allowing its use by people from different areas. This solution also allows AI developers to use their codes as a component of the flow, in EYeFlow, making it possible to use this component as part of the tool, which guarantees the use of all available facilities enabling a practical and agile delivery

How it works?

Flow

Intuitive interface for assembling processing flows, defines step by step the steps to be performed. The Flow is built using the components previously created using own and third party components for data processing. The interface features Drag And Drop technology, making it easy to create the flow. This interface provides components to classify, detect, perform OCR, perform analytical processes of people, measurement, among others.

Upload

Interface available in the tool to load the video for training the Neural Network, providing frames taken from the video to build the Dataset

Dataset

Interface for selecting frames from the video uploaded in the previous step, parameterization and customization of attributes for training in the Neural Network,

Marking

Tool available within the Dataset section to mark areas of interest in the frame

Training

Section available in Dataset to submit training for the Neural Network, performing the steps defined in the flow created.

Test

Validates the result of the neural network, through the dashboard interface and through the execution of the loaded video, if the result is not satisfactory, the solution must be submitted to a new training, aiming at improving the learning of the neural network.

Publish

Provides Flow for use, the neural network is already trained, ready to be delivered for Operation

Where should I go now?

• To begin
• Examples