You are using an unsupported browser. Please update your browser to the latest version on or before July 31, 2020.
close
Home > Sighthound Video > Using Sighthound Video > The region or boundary search doesn't do what I expect
The region or boundary search doesn't do what I expect
print icon

The region or boundary search doesn't do what I expect

 

If you have moved your camera, the region or boundary you originally drew may be in a different position than you expect. In the Search view you can select Show Region Zones from the View menu to display the region or boundary of the selected rule in red, as well as the center and bottom center of the moving object. This helps you understand why clips are included in your region rules.

 

When you are editing rules, you can click the arrows above the image to display images taken at earlier times. A current image from that camera is displayed on the right. Above it the text "Image From:" and a time range is displayed. This indicates an hour range from which the still image was taken. If the image from the current time is too dark to use, for example, you can click the arrows to the left of the time range to display an image from the previous hour. Keep going until you find an image that is suitable.

 

If you have not moved your camera, the following is a description of how regions work, and how they may behave differently than you expect.

 

Humans instinctively make 3D interpretations of 2D video images, you can look at an image in a video and (usually) interpret when a person is in front of something, or on top of something, etc. Currently, the application uses the center point of the bounding box around the object as the point that determines whether an object is inside a region, crossing a boundary, etc. As a result, cases that might create problems include the following:

 

  • You may see person's hand or leg cross into the region you drew, but the center point may not have actually crossed the line or box that you drew.
  • If a persons stands still and lifts one arm, making the bounding box around that person wider, the center will appear to have moved in that direction,
  • Dark shadows cast by a person are sometimes seen as part of the person (resulting in a bigger box). Or if a part of a person is in a dark shadow cast by another object, the box may be drawn around the portion of the person that is in the light (resulting in a smaller box). Both cases will effect where the center point of the person is estimated to be.

 

 

  • Other times, reflections or dark shadows are sometimes identified as separate people. This can actually be desirable in some cases (e.g., the person isn't visible but you see their shadow on a wall).
  • Sudden light changes can make the boxes grow, sometimes to the full size of the screen. We are working to minimize this.
  • Objects in front of other objects may appear to "split up" the object. For example, if someone walks behind a tree, the left part of the person may be seen as a different object than the right part.
scroll to top icon