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

Salient object detection requires a comprehensive and scalable receptivefield to locate the visually significant objects in the image. Recently, theemergence of visual transformers and multi-branch modules has significantlyenhanced the ability of neural networks to perceive objects at differentscales. However, compared to the traditional backbone, the calculation processof transformers is time-consuming. Moreover, different branches of themulti-branch modules could cause the same error back propagation in eachtraining iteration, which is not conducive to extracting discriminativefeatures. To solve these problems, we propose a bilateral network based ontransformer and CNN to efficiently broaden local details and global semanticinformation simultaneously. Besides, a Multi-Head Boosting (MHB) strategy isproposed to enhance the specificity of different network branches. Bycalculating the errors of different prediction heads, each branch canseparately pay more attention to the pixels that other branches predictincorrectly. Moreover, Unlike multi-path parallel training, MHB randomlyselects one branch each time for gradient back propagation in a boosting way.Additionally, an Attention Feature Fusion Module (AF) is proposed to fuse twotypes of features according to respective characteristics. Comprehensiveexperiments on five benchmark datasets demonstrate that the proposed method canachieve a significant performance improvement compared with thestate-of-the-art methods.

Receptive Field Broadening and Boosting for Salient Object Detection