Rank-DETR for High Quality Object Detection

TL;DR

Rank-DETR introduces rank-oriented architecture and loss functions to improve the accuracy of object ranking and localization in DETR-based detectors, leading to higher quality detection results across various backbones.

Contribution

The paper proposes a novel rank-oriented design and loss function for DETR-based detectors, enhancing localization accuracy and ranking quality for better detection performance.

Findings

Achieves state-of-the-art COCO detection results with various backbones

Improves localization accuracy at high IoU thresholds

Reduces false positive rates in object detection

Abstract

Modern detection transformers (DETRs) use a set of object queries to predict a list of bounding boxes, sort them by their classification confidence scores, and select the top-ranked predictions as the final detection results for the given input image. A highly performant object detector requires accurate ranking for the bounding box predictions. For DETR-based detectors, the top-ranked bounding boxes suffer from less accurate localization quality due to the misalignment between classification scores and localization accuracy, thus impeding the construction of high-quality detectors. In this work, we introduce a simple and highly performant DETR-based object detector by proposing a series of rank-oriented designs, combinedly called Rank-DETR. Our key contributions include: (i) a rank-oriented architecture design that can prompt positive predictions and suppress the negative ones to ensure lower false positive rates, as well as (ii) a rank-oriented loss function and matching cost design that prioritizes predictions of more accurate localization accuracy during ranking to boost the AP under high IoU thresholds. We apply our method to improve the recent SOTA methods (e.g., H-DETR and DINO-DETR) and report strong COCO object detection results when using different backbones such as ResNet-$50$, Swin-T, and Swin-L, demonstrating the effectiveness of our approach. Code is available at \url{https://github.com/LeapLabTHU/Rank-DETR}.