Skip to main content

Detection Regions and Object Size

Detection Regions​

Detection regions allow you to tell EyesOnIt what parts of your image or video to focus on for object detection. Detection regions are defined as rectangles and you can tell EyesOnIt to look at one or more regions. In the EyesOnIt web user interface (UI), regions can be defined by simply dragging the mouse on an image to create a rectangle. Let’s look at an example.

Let’s say I have a warehouse and I want to monitor vehicles entering the driveway. I have a camera mounted on the warehouse with a good view of the driveway. Within the EyesOnIt web UI, I can click on my image and draw a rectangle around the detection region. I can resize the rectangle by dragging the handles on the corner and the sides of the image. Creating this rectangle tells EyesOnIt to detect objects within the rectangular area. You can add more detection areas by adding more rectangles.

EyesOnIt detection region

Object Size​

In the image above, you may notice the green square in the upper left corner of the image. This green square indicates the estimated size of the object to detect. You can drag the corners of that green to change the size. You can also change the size with the Estimated Object Size slider below the image.

The green object size square can be moved to a specific location within your image. This is sometimes helpful as you resize the object size square to match the size of an object in your image. However, the location of the object size square is not used by EyesOnIt for processing. The only thing EyesOnIt cares about is the size of the object size rectangle.

The image below shows the detection region and object size rectangle for our driveway camera view.

EyesOnIt detection region and object size

The next section will explain how to create object descriptions that accurately detect vehicles in the driveway.