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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Hashing methods have been widely used for efficient similarity retrieval onlarge scale image database. Traditional hashing methods learn hash functions togenerate binary codes from hand-crafted features, which achieve limitedaccuracy since the hand-crafted features cannot optimally represent the imagecontent and preserve the semantic similarity. Recently, several deep hashingmethods have shown better performance because the deep architectures generatemore discriminative feature representations. However, these deep hashingmethods are mainly designed for supervised scenarios, which only exploit thesemantic similarity information, but ignore the underlying data structures. Inthis paper, we propose the semi-supervised deep hashing (SSDH) approach, toperform more effective hash function learning by simultaneously preservingsemantic similarity and underlying data structures. The main contributions areas follows: (1) We propose a semi-supervised loss to jointly minimize theempirical error on labeled data, as well as the embedding error on both labeledand unlabeled data, which can preserve the semantic similarity and capture themeaningful neighbors on the underlying data structures for effective hashing.(2) A semi-supervised deep hashing network is designed to extensively exploitboth labeled and unlabeled data, in which we propose an online graphconstruction method to benefit from the evolving deep features during trainingto better capture semantic neighbors. To the best of our knowledge, theproposed deep network is the first deep hashing method that can perform hashcode learning and feature learning simultaneously in a semi-supervised fashion.Experimental results on 5 widely-used datasets show that our proposed approachoutperforms the state-of-the-art hashing methods.

SSDH: Semi-supervised Deep Hashing for Large Scale Image Retrieval