utilising the Condensation algorithm. A number of tests are conducted and it is concluded that bootstrapped tracking is a promising approach when solving some of the inherent problems in model-based tracking

The problem of tracking curves in dense visual clutter is challenging. Kalman filtering is inadequate because it is based on Gaussian densities which, being unimodal, cannot represent simultaneous alternative hypotheses. The Condensation algorithm uses "factored sampling", previously

Particle filter methods based on color distribution can be used to track non-rigid moving objects in color videos. They are robust in case of noise or partial occlusions. However, using particle filters on color videos is sensitive to changes in the lighting conditions of the scene. The use of thermal infrared image sequences can help the tracking process, as thermal infrared imagery is not sensitive to lighting conditions. This paper presents a particle filter based method for object tracking using automatic cooperation between the color and the infrared modalities. As the infrared and the visible image sequences are acquired with different cameras, a pre-step is spatio-temporal registration. After spatio-temporal registration, the proposed method is able to continuously track the target despite difficult conditions appearing in one of the modality. Our cooperative tracking method is successfully applied on several experimental datasets. Different test sequences are presented, including tracking in the visible video with the help of the infrared modality, or tracking in the infrared with the help of the visible modality. Comments and future prospects raised by this method are finally

Some years ago a new tracker, the Condensation algorithm, came... In this paper an extension of the Condensation algorithm is introduced that relies on a single probability distribution to describe the likely states of multiple objects. By introducing an initialization density, observations can ow

The recognition of human gestures in image sequences is an important and challengingproblem that enables a host of human-computer interaction applications. This paper describes an incremental recognition strategy that is an extension of the “Condensation ” algorithm proposed by Isard and Blake

-based localization method based on the CONDENSATION algorithm [17, 18], a Bayesian filtering method that uses a samplingbased density representation. We show how the CONDENSATION algorithm can be used in a novel way to track the position of the camera platform rather than tracking an object in the scene. In addition

simultaneous alternative hypotheses. Extensions to the Kalman filter to handle multiple data associations work satisfactorily in the simple case of point targets, but do not extend naturally to continuous curves. A new, stochastic algorithm is proposed here, the Condensation algorithm --- Conditional

A new unsupervised filter-based feature selection method is introduced. Its principle consists in merging similar features into clusters using a distance measure derived from the correlation coefficient. Subsequently, only one representative feature is selected from each cluster. In experiments with real-world data, we show that the proposed method is benefical as a pre-filtering step for more sophisticated feature selection techniques. 1.

Abstract. Storing and using specific instances improves the performance of several supervised learning algorithms. These include algorithms that learn decision trees, classification rules, and distributed networks. However, no investigation has analyzed algorithms that use only specific instances

In an earlier paper, we introduced a new “boosting” algorithm called AdaBoost which, theoretically, can be used to significantly reduce the error of any learning algorithm that consistently generates classifiers whose performance is a little better than random guessing. We also introduced