Croatian Fish Dataset: Fine-grained classification of fish species in their natural habitat

Jonas Jäger, Marcel Simon, Joachim Denzler, Viviane Wolff, Klaus Fricke-Neuderth and Claudia Kruschel

Abstract

This paper presents a new dataset for fine-grained visual classification (FGVC) of fish species in their natural environment. It contains 794 images of 12 different fish species collected at the Adriatic sea in Croatia. All images show fishes in real live situations, recorded by high definition cameras. Remote and diver-based videography is used by a growing number of marine researchers to understand spatial and temporal variability of habitats and species, e.g. of fish assemblages. The required large numbers of independent observations necessitate the development of computer vision tools for an automated processing of high volumes of videos featuring high fish richness and density. As baseline experiment, we are using CNN features and a linear SVM classifier and achieve an accuracy of 67.55% on our dataset.

Session

Workshop: Machine Vision of Animals and their Behaviour (MVAB 2015)

Files

Paper (PDF, 3M)

DOI

10.5244/C.29.MVAB.6
https://dx.doi.org/10.5244/C.29.MVAB.6

Citation

Bibtex

@inproceedings{MVAB2015_6,
	title={Croatian Fish Dataset: Fine-grained classification of fish species in their natural habitat},
	author={Jonas Jäger and Marcel Simon and Joachim Denzler and Viviane Wolff and Klaus Fricke-Neuderth and Claudia Kruschel},
	year={2015},
	month={September},
	pages={6.1-6.7},
	articleno={6},
	numpages={7},
	booktitle={Proceedings of the Machine Vision of Animals and their Behaviour (MVAB)},
	publisher={BMVA Press},
	editor={T. Amaral, S. Matthews, T. Plötz, S. McKenna, and R. Fisher},
	doi={10.5244/C.29.MVAB.6},
	isbn={1-901725-57-X},
	url={https://dx.doi.org/10.5244/C.29.MVAB.6}
}