Adaptive approximation of the boundary surface of a neuron in confocal microscopy volumetric images.

In biomedical visualisation, the isosurface is usually used to represent (approximate) the boundary surface of the structure within biomedical volumetric images. However, in many confocal microscopy volumetric images of neurons, the grey values of the object and/or background are usually uneven. Therefore a fixed isosurface is not suitable for use in approximating the boundary surface of the neuron. A method is proposed to construct the adaptively approximating surface of the boundary surface of the neuron. In this method, the boundary surface of the neuron could be locally and adaptively approximated with different surface patches in different local regions. Consequently, the approximation accuracy has been considerably improved.