Feature tracking in real world scenes (or how to track a cow)

Abstract

In this paper we present a novel scheme for modelling and tracking complex real life objects. The scheme uses multiple models based on a variation of the Point Distribution Model [1] known as the Vector Distribution Model [2]. Inter and intra-class variation is separated using a variation on Linear Discriminant Analysis known as 'Delta Analysis'. The tracking scheme is stochastic and is based on modelling model characteristics by a set of discrete probability distributions, which are updated in an iterative manner. Initialisation is performed using low level processing and a predictor is used to initialise characteristic probabilities on subsequent frames. This scheme has been applied to the task of tracking livestock in a realistic farmyard situation.

Robust tracking and posture description for laboratory rodents using active shape models

“…These and related model-based techniques have been applied to a range of difficult image analysis problems, such as locating, interpreting, and measuring structures in medical images (Cootes et al, 1994;Redhead, Kotcheff, Taylor, Porter, & Hukins, 1997;Solloway, Hutchinson, Waterton, & Taylor, 1997), tracking and performing measurements on the nematode worm Caenorhabditis elegans (Butcher, Cootes, Courtney, Gill, & Lithgow, 1999), locating and tracking farm animals (Bulpitt, Boyle, & Forbes, 2000;Magee & Boyle, 1999;Marchant & Onyango, 1995;Sumpter & Bulpitt, 1998), and tracking pedestrians (Baumberg & Hogg, 1994). The important defining feature of such model-based approaches is that the model is first trained on a set of example images containing instances of the object of interest.…”