Objectness Score

Used to calculate the Confidence Score for Object Detection.

Objectness is essentially a measure of the probability that an object exists in a proposed region of interest, denoted P(Object).

During training, the model learns to output a high P(object) if there is an object in that anchor box (positive sample), and a low value otherwise (negative sample).

How does this work for YOLO?

• Idk, the anchors are a little complicated