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

Existing open-vocabulary object detectors typically enlarge their vocabularysizes by leveraging different forms of weak supervision. This helps generalizeto novel objects at inference. Two popular forms of weak-supervision used inopen-vocabulary detection (OVD) include pretrained CLIP model and image-levelsupervision. We note that both these modes of supervision are not optimallyaligned for the detection task: CLIP is trained with image-text pairs and lacksprecise localization of objects while the image-level supervision has been usedwith heuristics that do not accurately specify local object regions. In thiswork, we propose to address this problem by performing object-centric alignmentof the language embeddings from the CLIP model. Furthermore, we visually groundthe objects with only image-level supervision using a pseudo-labeling processthat provides high-quality object proposals and helps expand the vocabularyduring training. We establish a bridge between the above two object-alignmentstrategies via a novel weight transfer function that aggregates theircomplimentary strengths. In essence, the proposed model seeks to minimize thegap between object and image-centric representations in the OVD setting. On theCOCO benchmark, our proposed approach achieves 36.6 AP50 on novel classes, anabsolute 8.2 gain over the previous best performance. For LVIS, we surpass thestate-of-the-art ViLD model by 5.0 mask AP for rare categories and 3.4 overall.Code: https://github.com/hanoonaR/object-centric-ovd.

Bridging the Gap between Object and Image-level Representations for Open-Vocabulary Detection