Object detection of vehicles in images using the detr model

Abstract

The End-to-End Detection Transformer (DETR) is a novel object detection model introduced by researchers at Facebook AI Research (FAIR). DETR represents a departure from traditional object detection methods, which typically rely on region proposal networks (RPNs) and non-maximum suppression (NMS) for bounding box prediction. Instead, DETR employs a transformer architecture to perform object detection directly in end-to-end manner.