Configure Object Recognition for Best Results

To ensure the most accurate object recognition, you can set several parameters in the configuration file.

Specify a Minimum Size

If you expect objects (when present) to occupy a large proportion of the image, setting the MinSize parameter can reduce the number of false positives.

Perspective Effects

Objects may appear different to the training images you supplied, due to various transformations and perspective distortion. To specify the types of perspective effects that are expected, you can set the Geometry (for 2D objects) or Geometry3 (for 3D objects) parameters. The Geometry parameter defaults to HOMOGRAPHY, specifying that the object might be viewed at an angle, and allows for arbitrary focal length. This default setting is the most generally applicable, and is normally the correct setting when the object is being viewed through a camera (for example, the board of sponsors’ logos that often appears in post-match interviews at sports events). However, for certain applications a 2-D geometry (similarity transform) is more appropriate; for example, to detect a printed logo in a text document or scanned picture, or a broadcaster’s logo overlaid onto a video stream. For these cases, you can achieve much better false positive rates by setting Geometry to SIM2. Media Server still allows for scaling, translation, and rotation in the plane, but does not allow tilting of the camera axis.