No abstract

No abstract

No abstract

Arbitrarily shaped video objects are an important concept in modern video coding methods. The techniques presently used are not based on image elements but rather video objects having an arbitrary shape. The contour of this shape has to be transmitted and therefore compressed as well. The example is standard MPEG-4. In mobile video communication, transmission errors and loss of data can occur. In the paper, different approaches to object contour error evaluation are discussed. The contour error occurs as a result of approximation during encoding and transmission errors. As a compare criterion of the evaluation, Hausdorff distance between convex polygons is used. New evaluation is constructed on the technique of transforming the Cartesian x and y contour coordinate functions to the domain of polar coordinates yielding an r and a ij function. New evaluation respects specific local artifacts or errors better if compared to systems based on distance of point sets.

A variety of complex techniques, such as forward error correction (FEC), automatic repeat request (ARQ), hybrid ARQ or cross-layer optimization, require in their design and optimization phase a realistic model of binary error process present in a specific digital channel. Past and more recent modeling approaches focus on capturing one or more stochastic characteristics with precision sufficient for the desired model application, thereby applying concepts and methods severely limiting the model applicability (eg in the form of modeled process prerequisite expectations). The proposed novel concept utilizing a Vector Quantization (VQ)-based approach to binary process modeling offers a viable alternative capable of superior modeling of most commonly observed small-and large-scale stochastic characteristics of a binary error process on the digital channel. Precision of the proposed model was verified using multiple statistical distances against the data captured in a wireless sensor network logical channel trace. Furthermore, the Pearson's goodness of fit test of all model variants' output was performed to conclusively demonstrate usability of the model for realistic captured binary error process. Finally, the presented results prove the proposed model applicability and its ability to far surpass the capabilities of the reference Elliot's model. K e y w o r d s: binary error; VQ, wireless channel, logical channel, error model, wireless sensor network