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

Effective conservation actions require effective population monitoring.However, accurately counting animals in the wild to inform conservationdecision-making is difficult. Monitoring populations through image sampling hasmade data collection cheaper, wide-reaching and less intrusive but created aneed to process and analyse this data efficiently. Counting animals from suchdata is challenging, particularly when densely packed in noisy images.Attempting this manually is slow and expensive, while traditional computervision methods are limited in their generalisability. Deep learning is thestate-of-the-art method for many computer vision tasks, but it has yet to beproperly explored to count animals. To this end, we employ deep learning, witha density-based regression approach, to count fish in low-resolution sonarimages. We introduce a large dataset of sonar videos, deployed to record wildmullet schools (Mugil liza), with a subset of 500 labelled images. We utiliseabundant unlabelled data in a self-supervised task to improve the supervisedcounting task. For the first time in this context, by introducing uncertaintyquantification, we improve model training and provide an accompanying measureof prediction uncertainty for more informed biological decision-making.Finally, we demonstrate the generalisability of our proposed counting frameworkthrough testing it on a recent benchmark dataset of high-resolution annotatedunderwater images from varying habitats (DeepFish). From experiments on bothcontrasting datasets, we demonstrate our network outperforms the few other deeplearning models implemented for solving this task. By providing an open-sourceframework along with training data, our study puts forth an efficient deeplearning template for crowd counting aquatic animals thereby contributingeffective methods to assess natural populations from the ever-increasing visualdata.

Deep learning with self-supervision and uncertainty regularization to count fish in underwater images