Students' Research Circle    
 
 
2023
2022
2021
2020
2019
2018
2017
2016
2015
2014
2013
2012
2011
2010
2009
» 2008
The conference
Session 1
Jury 1
Session 2
Jury 2
Sponsors
Awards-list
2007
2006
2005
2004
2003
2002
Home » Archive » 2008

TDK conference 2008

Classification of horses based on the static investigation of the extremities
Sótonyi Kata - year 5
Dept. of Biomechanics, Semmelweis University; Large Animal Clinic, Üllő
Supervisors: Dr. Csende Zsolt, Dr. Szabó Ferenc

Abstract:

The quality of individual animals in horse sports is defined by the character, willingness, leg-structure, ride-ability, quality of walking-modes and jumping ability. Accumulating evidence suggest, that the above characteristics can be inherited, and they can be objectively measured (Bade et al, 1975). Therefore, recently more and more efforts have been made to develop a pre-selection method that would enable classification of animals based on the above characteristics. Beyond scientific reasons, it is also economically challenging to find a method, which can be applied to horses, so they can be selected efficiently, objectively and as early as possible.

In the present study we compared two breeding-stocks - either used for racing (n=13) or for hobby purposes (n=11) - based on their natural standing position. Using images of horses taken from the front, back, right and left sides of animals, we categorized individuals into a ranking system we developed, based on the classification system of Goody (2000). With the help of this custom made ranking system we analyzed the differences between breed-lines and individuals within a breeding stock, based on cluster analysis of the position of tendons around joints and the absolute number of muscles.

Our results revealed, that individuals can be divided in to three well defined clusters, with centroid distances of C1vsC2=3.61, C1vsC3=3.2 and C2vsC3=1.53. Individuals within a cluster can be well distinguished from each other, and they form a homogenous group within a given cluster. Based on tree clustering, also taking the Euclidean-distances into account, individuals can be ranked reliably.

Using our novel approach we were able to rank individual horses with a simply applicable method, therefore differentiating between the qualities of the breeding stocks and thus between their value. Ranking individuals within a stock is a further advantage of our method.



List of lectures