AlignDet: Aligning Pre-training and Fine-tuning in Object Detection

TL;DR

AlignDet introduces a unified pre-training framework that reduces discrepancies between pre-training and fine-tuning in object detection, leading to improved performance and generalization across various detectors and settings.

Contribution

It proposes a two-stage pre-training method, decoupling image and box domain training, to better initialize detectors and enhance downstream detection performance.

Findings

AlignDet improves FCOS by 5.3 mAP

AlignDet enhances RetinaNet by 2.1 mAP

AlignDet boosts Faster R-CNN by 3.3 mAP

Abstract

The paradigm of large-scale pre-training followed by downstream fine-tuning has been widely employed in various object detection algorithms. In this paper, we reveal discrepancies in data, model, and task between the pre-training and fine-tuning procedure in existing practices, which implicitly limit the detector's performance, generalization ability, and convergence speed. To this end, we propose AlignDet, a unified pre-training framework that can be adapted to various existing detectors to alleviate the discrepancies. AlignDet decouples the pre-training process into two stages, i.e., image-domain and box-domain pre-training. The image-domain pre-training optimizes the detection backbone to capture holistic visual abstraction, and box-domain pre-training learns instance-level semantics and task-aware concepts to initialize the parts out of the backbone. By incorporating the self-supervised pre-trained backbones, we can pre-train all modules for various detectors in an unsupervised paradigm. As depicted in Figure 1, extensive experiments demonstrate that AlignDet can achieve significant improvements across diverse protocols, such as detection algorithm, model backbone, data setting, and training schedule. For example, AlignDet improves FCOS by 5.3 mAP, RetinaNet by 2.1 mAP, Faster R-CNN by 3.3 mAP, and DETR by 2.3 mAP under fewer epochs.