Simple Multi-dataset Detection

Simple Multi-dataset Detection
复制标题

DOI:
10.1109/cvpr52688.2022.00742
复制
发表时间:
2021-02
期刊:
2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
影响因子:
--
通讯作者:
Xingyi Zhou;V. Koltun;Philipp Krähenbühl
Xingyi Zhou;V. Koltun;Philipp Krähenbühl
中科院分区:
其他
文献类型:
--
作者:
Xingyi Zhou;V. Koltun;Philipp Krähenbühl

文献摘要

被引文献

相似文献

如何构建一个通用的、宽泛的目标检测系统?我们使用所有曾经注释过的概念的所有标签。这些标签跨越具有潜在不一致分类的不同数据集。在本文中,我们提出了一种在多个大规模数据集上训练统一检测器的简单方法。我们使用特定于数据集的训练协议和LASS,但与特定于数据集的输出共享通用检测架构。我们将展示如何将这些特定于数据集的输出自动集成到一个通用的语义分类中。与以前的工作不同,我们的方法不需要手动进行分类协调。实验表明,在所有的数据集中,我们的学习分类法都优于专家设计的分类法。我们的多数据集检测器在每个训练域上的性能与数据集特定的模型一样好,并且可以推广到新的未知数据集,而不需要对它们进行微调。代码可在https://github.com/xingyizhou/UniDet.上找到
How do we build a general and broad object detection system? We use all labels of all concepts ever annotated. These labels span diverse datasets with potentially inconsistent taxonomies. In this paper, we present a simple method for training a unified detector on multiple large-scale datasets. We use dataset-specific training protocols and losses, but share a common detection architecture with dataset-specific outputs. We show how to automatically integrate these dataset-specific outputs into a common semantic taxonomy. In contrast to prior work, our approach does not require manual taxonomy reconciliation. Experiments show our learned taxonomy outperforms a expert-designed taxonomy in all datasets. Our multi-dataset detector performs as well as dataset-specific models on each training domain, and can generalize to new unseen dataset without fine-tuning on them. Code is available at https://github.com/xingyizhou/UniDet.