Compensation of binarized CLSM images for extraction of the form of a single neuron in the silkworm moth brain

This paper proposes a method for automatic extraction of the form of a single neuron from image series which are captured with confocal laser scanning microscopic (CLSM). Deficiency parts appear in binary images through extraction process because of noises in CLSM images. It is needed manual operation to compensate it. Our method connects some discrete points picked up from the presumable area which exist neuron with smooth curves. Using this method, the form of a single neuron is represented by smooth curves, and it is reduced that the burden of manual extraction processes. We applied our method the task of extraction of the form of a pre motor neuron in order to verify availability of the method.