In this paper we propose a general, object-oriented software architecture for model-based visual tracking. The library is general purpose with respect to object model, estimated pose parameters, visual modalities employed, number of cameras and objects, and tracking methodology. The base class structure provides the necessary building blocks for implementing a wide variety of both known and novel tracking systems, integrating different visual modalities, like as color, motion, edge maps etc., in a multi-level fashion, ranging from pixel-level segmentation, up to local features matching and maximum-likelihood object pose estimation. The proposed structure allows integrating known data association algorithms for simultaneous, multiple object tracking tasks, as well as data fusion techniques for robust, multi-sensor tracking; within these contexts, parallelization of each tracking algorithm can as well be easily accomplished. Application of the proposed architecture is demonstrated through the definition and practical implementation of several tasks, all specified in terms of a self-contained description language.

We present results of an investigation into the visualization and measurement of retinal blood flow distribution by means of singleexposure laser-speckle fundus photography. The technique relies on the speckle effect produced when laser light is scattered at a diffusing structure and on the fact that the speckle will be averaged out to some extent when the structure concerned is moving and /or decorrelating. We discuss two alternative techniques used to process the specklegrams obtained from the retina. The first technique uses an analog optical spatial filtering procedure to enhance the resulting variations in speckle contrast. Although first results have shown the basic usefulness of this technique, it suffers from fundamental disadvantages. In the second technique we digitize the specklegram and use digital image processing techniques to (1) convert contrast variations in the fundus photography into color variations and (2) obtain the blood flow of the vessels with a reasonably low statistical error. Subject terms: speckle; metrology; retinal blood flow; digital image processing; flow visualization; flow velocity.Optical Engineering 25 (6). 731 -735 (June 1986).

Abstract-In this paper, it is shown that synthetic images can be used to test specific use cases of a lane tracking algorithm which has been developed by Audi AG. This was achieved by setting up a highly configurable and extendable simulation framework "Virtual Test Drive". The main components are a traffic simulation, visualization and a sensor model which supplies ground truth data about the street lanes. Additionally, the visualization is used to generate synthetic camera sensor data. The testbed also contains a realistic driving dynamics simulation and a real image processing soft ECU (which is represented as a standard PC in the early development stages). One of the modules on the image processing ECU is a lane tracking algorithm. The algorithm is designed to calculate the transition curves while driving. This information can be used as input for driving assistance functions, e.g. lane departure warning. By running the lane tracker on a synthetic image it is possible to compare the results of the lane tracker with the ground truth data provided by the simulation. In this particular case, the information has been used to test and optimize parts of the systems by using specific and determined scenarios in the simulation.

Automotive manufacturers and customers wish to have fully automated driving functionality available in a huge set of locations, scenarios, and markets. This raises the need for universally applicable scene understanding and motion planning algorithms that do not rely on highly accurate maps or excessive infrastructure communication. In this paper we introduce two novel approaches for extracting a topological roadgraph with possible intersection options from sensor data along with a geometric representation of the available maneuvering space. Also, a search and optimization-based path planning method for guiding the vehicle along a selected track in the roadgraph and within the free-space is presented. We compare the methods presented in simulation and show results of a test drive with a research vehicle. Our evaluations show the applicability in low speed maneuvering scenarios and the stability of the algorithms even for low quality input data.

No abstract