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

Retrieving tracked-vehicles by natural language descriptions plays a criticalrole in smart city construction. It aims to find the best match for the giventexts from a set of tracked vehicles in surveillance videos. Existing worksgenerally solve it by a dual-stream framework, which consists of a textencoder, a visual encoder and a cross-modal loss function. Although someprogress has been made, they failed to fully exploit the information at variouslevels of granularity. To tackle this issue, we propose a novel framework forthe natural language-based vehicle retrieval task, OMG, which Observes MultipleGranularities with respect to visual representation, textual representation andobjective functions. For the visual representation, target features, contextfeatures and motion features are encoded separately. For the textualrepresentation, one global embedding, three local embeddings and a color-typeprompt embedding are extracted to represent various granularities of semanticfeatures. Finally, the overall framework is optimized by a cross-modalmulti-granularity contrastive loss function. Experiments demonstrate theeffectiveness of our method. Our OMG significantly outperforms all previousmethods and ranks the 9th on the 6th AI City Challenge Track2. The codes areavailable at https://github.com/dyhBUPT/OMG.

OMG: Observe Multiple Granularities for Natural Language-Based Vehicle Retrieval