Data Integration with Other Platforms

How is ReefCloud different to “CoralNet”?

ReefCloud and CoralNet share some of the same functionalities in terms of machine learning – in fact Oscar Beijbom who developed CoralNet has worked closely with the ReefCloud Project for years and is a friend to and advisor of ReefCloud. Building on this knowledge there are some minor improvements in the machine learning that ReefCloud applies which make the processing a little faster.

Where ReefCloud really differs is the integration of everybody’s datasets into a single data layer that organisations like the Global Coral Reef Monitoring Network (GCRMN) can use to improve insights into reef health at regional scales, as part of the Public Dashboard.

We know many users also have existing monitoring data in other benthic image analysis programs such as CoralNet and CPCe so we have worked to develop a bulk import function that offers the option to upload photos with the label codes used for point annotations, 'pre-training' the machine without having to re-do annotations in ReefCloud too!

Integration with other platforms

Commonly used tools for manual annotation of coral reef photos include Coral Point Count with Excel extensions (CPCe), PhotoQuad, PointCount99 and BIIGLE. These platforms use the same methodology as ReefCloud but don't offer semi-automatic or automatic analysis, all points need to be manually annotated. Other comparable platforms related to automatic analysis of coral include Squidle, CoralNet and MAIA, BIIGLE's extension.