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/

Zendy Plus

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

Zendy Tools

Zendy Open

Learning representative, robust and discriminative information from images isessential for effective person re-identification (Re-Id). In this paper, wepropose a compound approach for end-to-end discriminative deep feature learningfor person Re-Id based on both body and hand images. We carefully design theLocal-Aware Global Attention Network (LAGA-Net), a multi-branch deep networkarchitecture consisting of one branch for spatial attention, one branch forchannel attention, one branch for global feature representations and anotherbranch for local feature representations. The attention branches focus on therelevant features of the image while suppressing the irrelevant backgrounds. Inorder to overcome the weakness of the attention mechanisms, equivariant topixel shuffling, we integrate relative positional encodings into the spatialattention module to capture the spatial positions of pixels. The global branchintends to preserve the global context or structural information. For the thelocal branch, which intends to capture the fine-grained information, we performuniform partitioning to generate stripes on the conv-layer horizontally. Weretrieve the parts by conducting a soft partition without explicitlypartitioning the images or requiring external cues such as pose estimation. Aset of ablation study shows that each component contributes to the increasedperformance of the LAGA-Net. Extensive evaluations on four popular body-basedperson Re-Id benchmarks and two publicly available hand datasets demonstratethat our proposed method consistently outperforms existing state-of-the-artmethods.

Local-Aware Global Attention Network for Person Re-Identification Based on Body and Hand Images