Dynamic Sparse R-CNN

TL;DR

Dynamic Sparse R-CNN enhances the original Sparse R-CNN by introducing dynamic label assignment and proposal generation, leading to improved object detection performance and state-of-the-art results on COCO.

Contribution

The paper proposes dynamic label assignment and dynamic proposal generation to improve Sparse R-CNN's accuracy and adaptability during training and inference.

Findings

Achieves 47.2% AP on COCO 2017 validation set.

Surpasses Sparse R-CNN by 2.2% AP with ResNet-50 backbone.

Demonstrates effectiveness of dynamic designs in object detection.

Abstract

Sparse R-CNN is a recent strong object detection baseline by set prediction on sparse, learnable proposal boxes and proposal features. In this work, we propose to improve Sparse R-CNN with two dynamic designs. First, Sparse R-CNN adopts a one-to-one label assignment scheme, where the Hungarian algorithm is applied to match only one positive sample for each ground truth. Such one-to-one assignment may not be optimal for the matching between the learned proposal boxes and ground truths. To address this problem, we propose dynamic label assignment (DLA) based on the optimal transport algorithm to assign increasing positive samples in the iterative training stages of Sparse R-CNN. We constrain the matching to be gradually looser in the sequential stages as the later stage produces the refined proposals with improved precision. Second, the learned proposal boxes and features remain fixed for different images in the inference process of Sparse R-CNN. Motivated by dynamic convolution, we propose dynamic proposal generation (DPG) to assemble multiple proposal experts dynamically for providing better initial proposal boxes and features for the consecutive training stages. DPG thereby can derive sample-dependent proposal boxes and features for inference. Experiments demonstrate that our method, named Dynamic Sparse R-CNN, can boost the strong Sparse R-CNN baseline with different backbones for object detection. Particularly, Dynamic Sparse R-CNN reaches the state-of-the-art 47.2% AP on the COCO 2017 validation set, surpassing Sparse R-CNN by 2.2% AP with the same ResNet-50 backbone.