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

From human crowds to cells in tissue, the detection and efficient tracking ofmultiple objects in dense configurations is an important and unsolved problem.In the past, limitations of image analysis have restricted studies of densegroups to tracking a single or subset of marked individuals, or tocoarse-grained group-level dynamics, all of which yield incomplete information.Here, we combine convolutional neural networks (CNNs) with the modelenvironment of a honeybee hive to automatically recognize all individuals in adense group from raw image data. We create new, adapted individual labeling anduse the segmentation architecture U-Net with a loss function dependent on bothobject identity and orientation. We additionally exploit temporal regularitiesof the video recording in a recurrent manner and achieve near human-levelperformance while reducing the network size by 94% compared to the originalU-Net architecture. Given our novel application of CNNs, we generate extensiveproblem-specific image data in which labeled examples are produced through acustom interface with Amazon Mechanical Turk. This dataset contains over375,000 labeled bee instances across 720 video frames at 2 FPS, representing anextensive resource for the development and testing of tracking methods. Wecorrectly detect 96% of individuals with a location error of ~7% of a typicalbody dimension, and orientation error of 12 degrees, approximating thevariability of human raters. Our results provide an important step towardsefficient image-based dense object tracking by allowing for the accuratedetermination of object location and orientation across time-series image dataefficiently within one network architecture.

Towards dense object tracking in a 2D honeybee hive