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 state-of-the-art crowd counting algorithms rely excessively onlocation-level annotations, which are burdensome to acquire. When onlycount-level (weak) supervisory signals are available, it is arduous anderror-prone to regress total counts due to the lack of explicit spatialconstraints. To address this issue, a novel and efficient counter (referred toas CrowdMLP) is presented, which probes into modelling global dependencies ofembeddings and regressing total counts by devising a multi-granularity MLPregressor. In specific, a locally-focused pre-trained frontend is cascaded toextract crude feature maps with intrinsic spatial cues, which prevent the modelfrom collapsing into trivial outcomes. The crude embeddings, along with rawcrowd scenes, are tokenized at different granularity levels. Themulti-granularity MLP then proceeds to mix tokens at the dimensions ofcardinality, channel, and spatial for mining global information. An effectiveproxy task, namely Split-Counting, is also proposed to evade the barrier oflimited samples and the shortage of spatial hints in a self-supervised manner.Extensive experiments demonstrate that CrowdMLP significantly outperformsexisting weakly-supervised counting algorithms and performs on par withstate-of-the-art location-level supervised approaches.

CrowdMLP: Weakly-Supervised Crowd Counting via Multi-Granularity MLP