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

This paper presents a novel approach to Zero-Shot Action Recognition. Recentworks have explored the detection and classification of objects to obtainsemantic information from videos with remarkable performance. Inspired by them,we propose using video captioning methods to extract semantic information aboutobjects, scenes, humans, and their relationships. To the best of our knowledge,this is the first work to represent both videos and labels with descriptivesentences. More specifically, we represent videos using sentences generated viavideo captioning methods and classes using sentences extracted from documentsacquired through search engines on the Internet. Using these representations,we build a shared semantic space employing BERT-based embedders pre-trained inthe paraphrasing task on multiple text datasets. The projection of both visualand semantic information onto this space is straightforward, as they aresentences, enabling classification using the nearest neighbor rule. Wedemonstrate that representing videos and labels with sentences alleviates thedomain adaptation problem. Additionally, we show that word vectors areunsuitable for building the semantic embedding space of our descriptions. Ourmethod outperforms the state-of-the-art performance on the UCF101 dataset by3.3 p.p. in accuracy under the TruZe protocol and achieves competitive resultson both the UCF101 and HMDB51 datasets under the conventional protocol (0/50\%- training/testing split). Our code is available athttps://github.com/valterlej/zsarcap.

Tell me what you see: A zero-shot action recognition method based on natural language descriptions