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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Referring Image Segmentation (RIS) aims to segment target objects expressedin natural language within a scene at the pixel level. Various recent RISmodels have achieved state-of-the-art performance by generating contextualtokens to model multimodal features from pretrained encoders and effectivelyfusing them using transformer-based cross-modal attention. While these methodsmatch language features with image features to effectively identify likelytarget objects, they often struggle to correctly understand contextualinformation in complex and ambiguous sentences and scenes. To address thisissue, we propose a novel bidirectional token-masking autoencoder (BTMAE)inspired by the masked autoencoder (MAE). The proposed model learns the contextof image-to-language and language-to-image by reconstructing missing featuresin both image and language features at the token level. In other words, thisapproach involves mutually complementing across the features of images andlanguage, with a focus on enabling the network to understand interconnecteddeep contextual information between the two modalities. This learning methodenhances the robustness of RIS performance in complex sentences and scenes. OurBTMAE achieves state-of-the-art performance on three popular datasets, and wedemonstrate the effectiveness of the proposed method through various ablationstudies.

Synchronizing Vision and Language: Bidirectional Token-Masking AutoEncoder for Referring Image Segmentation