Scaling Open-Vocabulary Object Detection
This work addresses the data scarcity problem for open-vocabulary object detection, enabling web-scale training similar to image classification and language modeling, though it is incremental as it builds on existing self-training methods.
The paper tackles the limitation of training data in open-vocabulary object detection by scaling up detection data using self-training on web image-text pairs, resulting in OWLv2 and OWL-ST, which improve AP on LVIS rare classes from 31.2% to 44.6% (a 43% relative improvement).
Open-vocabulary object detection has benefited greatly from pretrained vision-language models, but is still limited by the amount of available detection training data. While detection training data can be expanded by using Web image-text pairs as weak supervision, this has not been done at scales comparable to image-level pretraining. Here, we scale up detection data with self-training, which uses an existing detector to generate pseudo-box annotations on image-text pairs. Major challenges in scaling self-training are the choice of label space, pseudo-annotation filtering, and training efficiency. We present the OWLv2 model and OWL-ST self-training recipe, which address these challenges. OWLv2 surpasses the performance of previous state-of-the-art open-vocabulary detectors already at comparable training scales (~10M examples). However, with OWL-ST, we can scale to over 1B examples, yielding further large improvement: With an L/14 architecture, OWL-ST improves AP on LVIS rare classes, for which the model has seen no human box annotations, from 31.2% to 44.6% (43% relative improvement). OWL-ST unlocks Web-scale training for open-world localization, similar to what has been seen for image classification and language modelling.