English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

The paradigm of large-scale pre-training followed by downstream fine-tuninghas been widely employed in various object detection algorithms. In this paper,we reveal discrepancies in data, model, and task between the pre-training andfine-tuning procedure in existing practices, which implicitly limit thedetector's performance, generalization ability, and convergence speed. To thisend, we propose AlignDet, a unified pre-training framework that can be adaptedto various existing detectors to alleviate the discrepancies. AlignDetdecouples the pre-training process into two stages, i.e., image-domain andbox-domain pre-training. The image-domain pre-training optimizes the detectionbackbone to capture holistic visual abstraction, and box-domain pre-traininglearns instance-level semantics and task-aware concepts to initialize the partsout of the backbone. By incorporating the self-supervised pre-trainedbackbones, we can pre-train all modules for various detectors in anunsupervised paradigm. As depicted in Figure 1, extensive experimentsdemonstrate that AlignDet can achieve significant improvements across diverseprotocols, such as detection algorithm, model backbone, data setting, andtraining schedule. For example, AlignDet improves FCOS by 5.3 mAP, RetinaNet by2.1 mAP, Faster R-CNN by 3.3 mAP, and DETR by 2.3 mAP under fewer epochs.

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