Adaptive detection for tracking moving biological objects in videomicroscopy sequences

摘要

We present a method to detect and track multiple moving biologicalobjects in images acquired by video microscopy. The automatic detectionis based upon the correlation of the image with a filter which variesadaptively to represent an object as it moves and deforms. The trackingis performed using a Kalman filter and a cost function which enable theposition of the moving objects to be predicted, refined and updated. Theefficiency of the method has been tested on real biological imagesequences and is illustrated by results obtained from the analysis oftypical biological video microscopy sequences