An exciting time for plankton imaging
by Jean-Olivier Irisson, on
Teams at OOV and RSMAS (and now OSU) have been working with instruments that allow us to take images of planktonic animals in situ: the Underwater Video Profiler (UVP) at OOV and the In Situ Ichthyoplankton Imaging System (ISIIS) at RSMAS/OSU. Half of the research done through this PUF project was built on those instruments, coupling them for the VISUFRONT cruise during the first year of the project. We worked on analysing the images and data recorded by these instruments since then.
My coming to RSMAS this week was an occasion to discuss plankton imaging again, two years later, with Bob Cowen. It was a long, interesting and fruitful discussion. It seems we are at a tipping point for plankton imaging. On the RSMAS/OSU side, ISIIS has now been used in several projects and multiple scientific papers have been and are being written. An automated image processing and segmenting routine is operational and fast. The planktonportal.org collaborative website for plankton identification generates a steady stream of human-validated identifications for images taken by ISIIS. The kaggle Data Science Bowl competition spurred several promising algorithms for automated image classification. On the OOV side, the UVP is being miniaturised to be included in autonomous underwater instruments and collect more data in more places. A web application for semi-automated, collaborative image classification by scientists has been developed and is ready for UVP and ISIIS images. Dr Ben Graham, one of the authors of a winning algorithms of the kaggle competition will be coming to OOV to discuss including his work as the classification backend of this application. And, on both sides of the Atlantic, the instruments and associated tools are becoming standard enough to be considered by environment monitoring programs.
So overall, the times are exciting and full of opportunities for all the teams involved in this PUF program!