Documentation for Axxon Next 4.4.6. Documentation for other versions of Axxon Next is available too.

Previous page General information on Neural Analytics  Detection tools implementation on processors Next page

To train your neural network, you have to collect (and submit to AxxonSoft) video recordings from your actual cameras with the same resolution, expected weather conditions and time of day as in your future application.

For example, if your neural network is intended to analyze outdoor video feeds, your footage must contain all range of weather conditions (sun, rain, snow, fog, etc.) in different times of day (daytime, twilight, night).

Extra requirements for video footage for each neural analytics tool are listed in the following table:

Tool

Requirements

Neural Filter

No less than 1000 frames containing objects of interest in given scene conditions, and the same amount of footage containing no objects (background footage).

Neural Tracker

3 to 5 minutes of video containing objects of interest in given scene conditions.

Posture detection tools

No less than 100 different persons in given scene conditions.

Attention! Different conditions mean, among others, different postures of an individual in scene (tilting, different limbs patterns, etc.).

Segmenting detector *

3 to 5 minutes of video containing objects of interest in given scene conditions.

Food recognition *

You need to submit no less than 80% of actual menu positions. Each position requires 20 to 40 images shot in different conditions. 

Note

* Will be available in future versions of Axxon Next software.

The requirements may be changed or added to at any time.

  • No labels