Behavier Analysis of House Dust Mites by Robust Difference Accumulation

Abstract

A method is proposed for estimating density of active mites in the media by processing microscopic mite images taken in successive order. An estimated background image of the scene has an important role for constructing a robust detector of object regions from the image sequence. One of the difficulty comes from the fact of co-occurence of media motion with mite motion. Then a mathematical model for calculating mite density is described, which utilize the area obtained by the detector and an estimator of response area corresponding to mite motion. The experimental results show the effectivity of the proposed method.