Pixels, Not Boxes: How a Detection Transformer Paints a Street

100 questions, four answers, each painted pixel by pixel. The hatched area is what a box would have counted as the object.* Left of the line, a box counts everything inside it, including the road. Right of the line, a mask keeps only the object’s own pixels. Eight animated steps, one street: how RF-DETR finds every object and paints its exact pixels, and where it struggles. Look at a photo of a street. A cyclist rides across a zebra crossing. One car waits before the crossing, another drives away, and a dog sits on the pavement. Ask a computer where the objects are, and most systems answer wit