Knowledge Base

Problem

When Deep Learning is selected as the Body-Point Detection technique in EthoVision XT, the Behavior Recognition module becomes grayed out and unavailable. Users cannot use both features in the same experiment.

Details

Affects EthoVision XT 19 (and likely nearby versions). When a user selects 'Deep Learning' under Body-Point Detection Technique in Experiment Settings, the Automatic Behavior Recognition option becomes grayed out and cannot be enabled. This is a documented software limitation — the two modules are architecturally incompatible within the same experiment at this time. It is not a configuration error or licensing issue.

Solution

1. Confirm the limitation: Deep Learning (body-point detection) and Automatic Behavior Recognition cannot be enabled simultaneously in the same experiment. This is a current software limitation, not a setup error.
2. If you need Behavior Recognition, switch the Body-Point Detection Technique from 'Deep Learning' to a standard method (e.g., Center-point, 3-point). Behavior Recognition will then become available.
3. If you need Deep Learning body-point detection, you will not be able to use Automatic Behavior Recognition in the same experiment. Consider running two separate experiments — one with Deep Learning for detailed posture/body-point data, and one with standard detection plus Behavior Recognition — on the same video files.
4. Monitor EthoVision XT release notes and Noldus communications for updates, as this limitation may be addressed in a future version.
5. Contact Noldus Support if you believe this behavior is not consistent with your licensed modules or version documentation.