Skip to main content

FLEX AI: Get started

  • Updated

Applies to: FLEX AI

Summary:

FLEX AI enables you to enhance camera capabilities, enabling them to detect and track previously unidentifiable objects. While many cameras can detect and track people and specific vehicle types, it is a different matter identifying shopping carts, forklift trucks, hovercrafts, or items on a conveyor belt.

With FLEX AI you can create custom object detection models for solid objects that can then be deployed to cameras, enabling you to detect objects that standard analytics does not already detect.

FLEX AI is a cloud-based application, and its results are compatible with deployment to P-series AI cameras.

How it Works:

FLEX AI operates on a subscription-based model, granting users comprehensive access to train, process, and download custom detection models for end customers. However, the allocation of licenses requires a careful consideration of each end customer, with a separate FLEX AI license needed for each one. This process is typically managed by a STEP Partner, who oversees the distribution of licenses to ensure effective deployment.

For actual deployment onto cameras, FLEX AI uses a perpetual licensing system, distinct from its subscription model. Each camera necessitates its own perpetual license to run FLEX AI models effectively. This approach ensures that users can deploy and utilize the models on their cameras without the constraints of a subscription-based licensing system.

In terms of training models, FLEX AI currently supports the MP4 format or tagged WAVE Sync videos for training. 

NOTE: Each project can only train one model.

Camera Compatibility:

Refer to the following article for camera compatibility: Which Cameras are Compatible With FLEX AI

Note: Each camera can only run one FLEX AI model, although it may run in parallel with WiseAI.

FLEX AI Flow:

The general FLEX AI process is as follows:

  1. Ensure you have a license to develop new models.
    Refer to: Licensing Hanwha Vision Cloud Products
  2. Sign in to Cloud Portal and select FLEX AI.
    Refer to: How do I launch Cloud Portal?

  3. Create a new project.
    Refer to: FLEX AI: How do I create and manage FLEX AI projects?

  4. Import training clips.
    Refer to: FLEX AI: How do I add a video clip for FLEX AI usage?

  5. Annotate the clips.
    Refer to: FLEX AI: How do I annotate objects?

  6. Send the model for training.

  7. Evaluate the model, and improve if necessary.
    Refer to: FLEX AI: How do I improve my object detection model?

  8. Download the model to your computer.

  9. Upload the model to a compatible camera.
    Refer to: FLEX AI: How do I upload a FLEX AI model to a camera?
  •  

Related articles