Identifying Individual Clown Fish

Xiao  Liu; Robert Fisher

Identifying Individual Clown Fish

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

This paper presents a clustering approach to estimating the number of individuals present in a set of Amphiprion clarkii (clownfish) observed live from a Taiwanese coral reef. Previous approaches to individual identification have addressed species where there are many individuals with some distinctive pattern and few observations of each. By contrast, fish observations from a fixed camera tend to produce many observations of a few individuals, whose differences in appearance are more quantitative than qualitative. The paper describes a k-means clustering approach to estimating the number of individuals. Although there is no absolute ground truth (e.g. from tagging),
our data analysis shows 6 or 7 individuals were present in a collection of 785 observations.

Original language	English
Title of host publication	Proceedings of Visual Observation and Analysis of Vertebrate And Insect Behavior 2016
Number of pages	4
Publication status	Published - 4 Dec 2016
Event	Visual Observation and Analysis of Vertebrate And Insect Behavior 2016 - Cancun, Mexico Duration: 4 Dec 2016 → 4 Dec 2016 http://www.icpr2016.org/site/

Conference

Conference	Visual Observation and Analysis of Vertebrate And Insect Behavior 2016
Abbreviated title	VAIB 2016
Country/Territory	Mexico
City	Cancun
Period	4/12/16 → 4/12/16
Internet address	http://www.icpr2016.org/site/

Access to Document

vaibliuAccepted author manuscript, 803 KB

http://homepages.inf.ed.ac.uk/rbf/VAIB16PAPERS/vaibliu.pdf

Fish4Knowledge: Fish4Knowledge: Supporting humans in knowledge gathering and question answering wrt marine and environmental monitoring through analysis of multiple video streams
Fisher, B. (Principal Investigator) & Chen-Burger, J. (Co-investigator)
EU government bodies
1/10/10 → 30/09/13
Project: Research
Fish4Knowledge: Fish4Knowledge: Supporting humans in knowledge gathering and question answering wrt marine and environmental monitoring through analysis of multiple video streams
Fisher, B. (Principal Investigator) & Chen-Burger, J. (Co-investigator)
EU government bodies
1/10/10 → 30/09/13
Project: Research

Cite this

@inproceedings{0dc0aa57670049dbab437fe002e07a13,

title = "Identifying Individual Clown Fish",

abstract = "This paper presents a clustering approach to estimating the number of individuals present in a set of Amphiprion clarkii (clownfish) observed live from a Taiwanese coral reef. Previous approaches to individual identification have addressed species where there are many individuals with some distinctive pattern and few observations of each. By contrast, fish observations from a fixed camera tend to produce many observations of a few individuals, whose differences in appearance are more quantitative than qualitative. The paper describes a k-means clustering approach to estimating the number of individuals. Although there is no absolute ground truth (e.g. from tagging),our data analysis shows 6 or 7 individuals were present in a collection of 785 observations.",

author = "Xiao Liu and Robert Fisher",

year = "2016",

month = dec,

day = "4",

language = "English",

booktitle = "Proceedings of Visual Observation and Analysis of Vertebrate And Insect Behavior 2016",

note = "Visual Observation and Analysis of Vertebrate And Insect Behavior 2016, VAIB 2016 ; Conference date: 04-12-2016 Through 04-12-2016",

url = "http://www.icpr2016.org/site/",

}

TY - GEN

T1 - Identifying Individual Clown Fish

AU - Liu, Xiao

AU - Fisher, Robert

PY - 2016/12/4

Y1 - 2016/12/4

N2 - This paper presents a clustering approach to estimating the number of individuals present in a set of Amphiprion clarkii (clownfish) observed live from a Taiwanese coral reef. Previous approaches to individual identification have addressed species where there are many individuals with some distinctive pattern and few observations of each. By contrast, fish observations from a fixed camera tend to produce many observations of a few individuals, whose differences in appearance are more quantitative than qualitative. The paper describes a k-means clustering approach to estimating the number of individuals. Although there is no absolute ground truth (e.g. from tagging),our data analysis shows 6 or 7 individuals were present in a collection of 785 observations.

AB - This paper presents a clustering approach to estimating the number of individuals present in a set of Amphiprion clarkii (clownfish) observed live from a Taiwanese coral reef. Previous approaches to individual identification have addressed species where there are many individuals with some distinctive pattern and few observations of each. By contrast, fish observations from a fixed camera tend to produce many observations of a few individuals, whose differences in appearance are more quantitative than qualitative. The paper describes a k-means clustering approach to estimating the number of individuals. Although there is no absolute ground truth (e.g. from tagging),our data analysis shows 6 or 7 individuals were present in a collection of 785 observations.

M3 - Conference contribution

BT - Proceedings of Visual Observation and Analysis of Vertebrate And Insect Behavior 2016

T2 - Visual Observation and Analysis of Vertebrate And Insect Behavior 2016

Y2 - 4 December 2016 through 4 December 2016

ER -

Identifying Individual Clown Fish

Abstract

Conference

Access to Document

Fingerprint

Projects

Fish4Knowledge: Fish4Knowledge: Supporting humans in knowledge gathering and question answering wrt marine and environmental monitoring through analysis of multiple video streams

Fish4Knowledge: Fish4Knowledge: Supporting humans in knowledge gathering and question answering wrt marine and environmental monitoring through analysis of multiple video streams

Cite this