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

With the development of deep neural networks, the performance of crowdcounting and pixel-wise density estimation are continually being refreshed.Despite this, there are still two challenging problems in this field: 1)current supervised learning needs a large amount of training data, butcollecting and annotating them is difficult; 2) existing methods can notgeneralize well to the unseen domain. A recently released synthetic crowddataset alleviates these two problems. However, the domain gap between thereal-world data and synthetic images decreases the models' performance. Toreduce the gap, in this paper, we propose a domain-adaptation-style crowdcounting method, which can effectively adapt the model from synthetic data tothe specific real-world scenes. It consists of Multi-level FeatureawareAdaptation (MFA) and Structured Density map Alignment (SDA). To be specific,MFA boosts the model to extract domain-invariant features from multiple layers.SDA guarantees the network outputs fine density maps with a reasonabledistribution on the real domain. Finally, we evaluate the proposed method onfour mainstream surveillance crowd datasets, Shanghai Tech Part B,WorldExpo'10, Mall and UCSD. Extensive experiments evidence that our approachoutperforms the state-of-the-art methods for the same cross-domain countingproblem.

Feature-aware Adaptation and Density Alignment for Crowd Counting in Video Surveillance