Topics

Range Image Segmentation by Curve Grouping


Abstract: 
A range image segmentation method based on a recursive adaptive regression model prediction for detecting range image step discontinuities which are present at object face borders. Detected face borders guides subsequent region growing step where neighbouring face curves are grouped together. Region growing based on curve segments instead of pixels like in classical approaches significantly speed up the algorithm. Curves to be grown are represented using the cubic spline model. Curves from the same region are required to have similar curvature and slope but they do not need to be of maximal length through the corresponding object face.

Virtual Reality

   Michal Haindl    Pavel Slavik    Stepan Kment    Ondrej Novak    Tomas Zalusky    Vaclav Blazek    Zdenek Pichlik

Abstract: 
The project studied techniques for automatic acquisition of complex virtual reality models from real image data (range and colour) together with virtual space navigation problems (for some of these techniques see corresponding specialized demonstrations). The results are demonstrated on the model of the Department of Modern Art of the National Gallery. This collection of images, drawings and statues from the period of 20th century is located in a functionalistic building in Prague. However due to the request of the National Gallery most of the exhibited art items are removed from this demonstration.

Texture Segmentation Using Recursive Markov Random Field Parameter Estimation


Abstract: 
An efficient and robust type of unsupervised colour texture segmentation method.

Fast Segmentation of Planar Surfaces in Range Images


Abstract: 
An algorithm for planar face segmentation in range images. The segmentation is based on a combination of recursive adaptive regression model prediction for detecting range image step discontinuities which are present at object face borders and of a region growing on surface lines. Border pixels are detected in two perpendicular directions and detection results are combined together. Two predictors in each direction use identical contextual information from the pixel's neighbourhood and they mutually compete for the most optimal discontinuity detection.