Existing work in the field of Multiple Instance Learning (MIL) have only looked at the standard two-class problem assuming both positive and negative bags are available. In this work, we propose the first analysis of the one-class version of MIL problem where one is only provided input data in the form of positive bags. We also propose an SVM-based formulation to solve this problem setting. To ...

This paper deals with a method for removing a ghost target that is not a real object from the output of a multiple object-tracking algorithm. This method uses an artificial neural network (multilayer perceptron) and introduces a structure, learning, verification, and evaluation method for the artificial neural network. The implemented system was tested at an intersection in a city center. Resul...

In four experiments we address the question whether several visual objects can be selected voluntarily (exogenously) and then tracked in a Multiple Object Tracking paradigm and, if so, whether the selection involves a different process. Experiment 1 showed that items can indeed be selected based on their labels. Experiment 2 showed that to select the complement set to a set that is automaticall...

10.

Space situational awareness requires the ability to initialize state estimation from short measurements and the reliable association of observations to support the characterization of the space environment. The electro-optical systems used to observe space objects cannot fully characterize the state of an object given a short, unobservable sequence of measurements. Further, it is difficult to a...

This paper proposes the set of target trajectories as the state variable for multiple target tracking. The main objective of multiple target tracking is to estimate an unknown number of target trajectories given a sequence of measurements. This quantity of interest is perfectly represented as a set of trajectories without the need of arbitrary parameters such as labels or ordering. We use finit...

This paper proposes an environment-dependent vehicle dynamic modelling approach considering interactions between the noisy control input of a dynamic model and the environment in order to make best use of domain knowledge. Based on this modelling, a new domain knowledge-aided moving horizon estimation (DMHE) method is proposed for ground moving target tracking. The proposed method incorporates ...

The random-set framework for multiple target tracking offers a distinct alternative to the traditional approach to multiple target tracking by treating the collections of individual targets and observations as finite-sets. The multi-target state is predicted and updated recursively based on the set-valued observation. The complexity of computing the multi-target recursion grows exponentially wi...

Based on our previous work we propose to track multiple ground targets with GMTI (GroundMoving Target Indicator) sensors as well as with imagery sensors. The scope of this paper is to fuse the attribute type information given by heterogeneous sensors with DSmT (Dezert Smarandache Theory) and to introduce the type results in the tracking process to improve its performances.

One of the major problems in multiple target tracking is to obtain an accurate association between targets and noisy measurements. We introduce a new scheme, called Constrained Optimal Data Association (CODA), that finds the optimal data association by a MAP estimation method and uses a new energy function. In this scheme, the natural constraints between targets and measurements are defined so ...