Estimation of Object Distribution Using Value Function for Crowded Swarm Analysis of Soldier Crab

Abstract

This paper proposes a new method to estimate the distribution of some similar objects that have no specific feature like color and shape. In the field of animal behavior analysis, the estimation of the crowded swarm distribution with automated way of image processing is necessary. However, to trace the individual objects those may overlap each other in video sequence is quite difficult. This paper proposes to introduce a value function effective for the crowded swarm analysis, and shows the affirmative results acquired for the Soldier Crab behavior analysis.