PoCo: an automated polyp counter

Project goals:

Manual annotation and counting of entities in underwater photographs is common in many branches of marine biology. With a marked increase of jellyfish populations worldwide, understanding the dynamics of the polyp (scyphistoma) stage of their life-cycle is becoming increasingly important. In-situ studies of polyp population dynamics are scarce due to small size of the polyps and tedious manual work required to annotate and count large numbers of items in underwater photographs.

The goal of this project is to develop automated algorithms for polyp census that will enable processing very large collections of polyp images, which will facilitate carrying out census at scales order of magnitude greater than what is achievable by manual labor.

Collaborating partners: University of Ljubljana, Faculty of Computer and Information Science; National Institute of Biology, Marine Biology Station

PoCo v1.0

The PoCo version 1.0 is based on a two-stage approach: polyp proposal detection and verification. The approach applies learning-based methodology that allows it to train on a large collection of annotated images. The algorithm can be trained for specific domain (i.e., photos taken under certain conditions) or it can be trained on a diverse set of images to achieve maximal generality.

PoCov1 sketch.PNG
Figure 1:Outline of PoCo v1.0 automated counter processing pipeline.

The paper describing the details of the method is currently under review. Our results indicate that the method accuracy is on par with expert human annotator when applied to a reasonably constrained domain on which it was trained.

PoCo v1.0 code and citation:

The PoCo v1.0 implementation is a mixture of Matlab/C++ libraries and is available here.

If you use PoCo in your project or publication, please cite the following paper:

@article{ VodopivecPoCo2018,
  Title = {Towards automated scyphistoma census in underwater imagery: a useful research and monitoring tool},
  Author = {Vodopivec, Martin and Mandeljc, Rok and Makovec, Tihomir and Malej, Alenka and Kristan, Matej},
  Booktitle = {Journal of Sea Research},
  Year = {2018}


  • izr. prof. dr. Matej Kristan (PI on FRI side)
  • dr. Martin Vodopivec (PI on MBS side)
  • prof. dr. Alenka Malej
  • Tihomir Makovec
  • dr. Rok Mandeljc
  • Vitjan Zavrtanik


Scientific output of the project is described in these publications:

Publications for the PoCo project