Hello,
Is it possible to make facial recognition function with visual proxies? I am unable to complete the facial recognition training since the viewer's menu is grayed out. My own experience has shown that, at least in the short view that the green box provided, the facial recognition capability appears to function. Is it safe to infer that this occurs when the image proxy is cropped? The green box is occasionally not centered on the face.
Is this a constraint or am I not doing this correctly?
Best regards,
Face recognition runs on each individual image. IMatch runs it for the master as well as the proxy.
Annotations are stored per image. When you look at the master in the Viewer, IMatch would load the annotations, including face annotations, for the master. But when you work with the proxies, you may see a cropped or resized or even rotated image instead, breaking annotation positions and sizes. This is why annotations are disabled when you enable the proxy display int he Viewer.
Thank you for your answer. :)
Quote from: Mario on December 26, 2023, 10:15:39 AMFace recognition runs on each individual image. IMatch runs it for the master as well as the proxy. eggy car (https://eggycar.net/)
Annotations are stored per image. When you look at the master in the Viewer, IMatch would load the annotations, including face annotations, for the master. But when you work with the proxies, you may see a cropped or resized or even rotated image instead, breaking annotation positions and sizes. This is why annotations are disabled when you enable the proxy display int he Viewer.
Thank you for your answer. :)