A hybrid system to detect hand orientation in stereo images

Adaptive color segmentation-a comparison of neural and statistical methods

“…The matrix performs the linear interpolation within the tessellation cell of node . The output of a single LLM network for an input feature vector is (1) The "winner" node is determined by finding the Voronoi cell that contains the input vector , i.e., must obey the minimality condition (2) This leads to the learning steps for a training sample…”

“…In a collaboration at Bielefeld University we investigated a hybrid system to combine the advantages of knowledgebased and neural methods [2], [8] that we applied to a Manuscript received December 18, 1995 challenging and important problem in pattern recognition, the recognition of human hands in real camera images. The solution of this task forms the basis for a number of subsequent processing steps during which further geometrical features, such as hand orientation and information on finger posture, are extracted.…”

Hybrid object recognition in image sequences

“…Contrarily to the work presented in [4,1] where the neural network serves as an interface for the instantiation of hybrid concepts now the holistic results are used to guide a decomposition process inside the semantic network to get structured object hypotheses efficiently. To increase the robustness of our analysis and to speedup processing we extended this approach to image sequences where on the basis of previous structural results the current image is analyzed.…”