Skip to Main Content
Though electron tomography opens up new possibilities in imaging the microtubule and the fine plus-end structures, the interpretation of the acquired data remains an obstacle due to the low SNR and the cluttered cellular environment. The automatic extraction of plus-end is especially challenging since they have complex and varying conformations beyond the capacity of existing segmentation methods. We propose an automated approach to extracting the microtubule plus-end with a coarse to fine scale scheme consisting of volume enhancement and plus-end segmentation. To make the segmentation robust against confusing image features, we have fully incorporated the prior knowledge of microtubules and plus-ends into our model-based framework. Experimental results demonstrate that our automated method produces results comparable to the manual segmentation but using only a fraction of the manual segmentation time. The automated approach also segments more fine structures that could be overlooked by human operators.