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

Archaeology has long faced fundamental issues of sampling and scalarrepresentation. Traditionally, the local-to-regional-scale views of settlementpatterns are produced through systematic pedestrian surveys. Recently,systematic manual survey of satellite and aerial imagery has enabled continuousdistributional views of archaeological phenomena at interregional scales.However, such 'brute force' manual imagery survey methods are both time- andlabor-intensive, as well as prone to inter-observer differences in sensitivityand specificity. The development of self-supervised learning methods offers ascalable learning scheme for locating archaeological features using unlabeledsatellite and historical aerial images. However, archaeological features aregenerally only visible in a very small proportion relative to the landscape,while the modern contrastive-supervised learning approach typically yields aninferior performance on highly imbalanced datasets. In this work, we propose aframework to address this long-tail problem. As opposed to the existingcontrastive learning approaches that treat the labelled and unlabeled dataseparately, our proposed method reforms the learning paradigm under asemi-supervised setting in order to utilize the precious annotated data (<7% inour setting). Specifically, the highly unbalanced nature of the data isemployed as the prior knowledge in order to form pseudo negative pairs byranking the similarities between unannotated image patches and annotated anchorimages. In this study, we used 95,358 unlabeled images and 5,830 labelledimages in order to solve the issues associated with detecting ancient buildingsfrom a long-tailed satellite image dataset. From the results, oursemi-supervised contrastive learning model achieved a promising testingbalanced accuracy of 79.0%, which is a 3.8% improvement as compared to otherstate-of-the-art approaches.

Semi-Supervised Contrastive Learning for Remote Sensing: Identifying Ancient Urbanization in the South Central Andes