Object Tracking with Adaptive Multicue Incremental Visual Tracker

Algorithm 1

Multicue based IVT algorithm.

Initialization

Locate the target manually in the first frame, and use a single particle to indicate this

location. Set the initial relative sharpness factors as for cues. Initialize the

eigenbasis to be empty, and the mean to be the appearance of the target in the first frame.

for  t  =  1  to  T

 (1) Spread the target states at time to time using the state dynamic model.

 (2) For each new state corresponding to particle at time , find its corresponding

    weight in feature space based on its likelihood under the observation models.

 (3) Based on each cue’s relative sharpness factor , . Combine multiple cues

    by calculating the new weight for each particle as .

 (4) Store the image window corresponding to the most likely particle. When the desired

    number of new images have been accumulated, perform an incremental update (with a

    forgetting factor) of the eigenbasis, mean, and effective number of observations.

 (5) Update the relative sharpness factor for each cue at time as based on the

    estimated target state and the particle distribution.

end  for