Featurized Query R-CNN

TL;DR

Featurized Query R-CNN introduces a query generation network to produce object queries within the Faster R-CNN framework, reducing computation and improving generalization, achieving superior speed-accuracy trade-offs on COCO.

Contribution

The paper proposes a novel featurized query generation method integrated into Faster R-CNN, addressing efficiency and generalization issues in query-based object detection.

Findings

Outperforms state-of-the-art R-CNN detectors on COCO

Reduces computational burden compared to multi-stage decoders

Enhances generalization of object queries

Abstract

The query mechanism introduced in the DETR method is changing the paradigm of object detection and recently there are many query-based methods have obtained strong object detection performance. However, the current query-based detection pipelines suffer from the following two issues. Firstly, multi-stage decoders are required to optimize the randomly initialized object queries, incurring a large computation burden. Secondly, the queries are fixed after training, leading to unsatisfying generalization capability. To remedy the above issues, we present featurized object queries predicted by a query generation network in the well-established Faster R-CNN framework and develop a Featurized Query R-CNN. Extensive experiments on the COCO dataset show that our Featurized Query R-CNN obtains the best speed-accuracy trade-off among all R-CNN detectors, including the recent state-of-the-art Sparse R-CNN detector. The code is available at {https://github.com/hustvl/Featurized-QueryRCNN.