Learning a unified label space

Xingyi Zhou, Vladlen Koltun, Philipp Kraehenbuehl

Keywords: object detection, image recognition, computer vision

Abstract: How do we build a general and broad object detection system? We use all labels of all concepts ever annotated. These labels span many diverse datasets with potentially inconsistent semantic labels. In this paper, we show how to integrate these datasets and their semantic taxonomies in a completely automated fashion. Once integrated, we train an off-the-shelf object detector on the union of the datasets. This unified recognition system performs as well as dataset-specific models on each training domain, but generalizes much better to new unseen domains. Entries based on the presented methodology ranked first in the object detection and instance segmentation tracks of the ECCV 2020 Robust Vision Challenge.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

One-sentence Summary: We learn to unify the taxonomy of different detection datasets and train a strong detector on all of them.

Reviewed Version (pdf): https://openreview.net/references/pdf?id=i8YbubLE2