Hough Transform-based Technique for Automated Carbon Nanocone Segmentation

Abstract

A new technique to automatically segment the L-shaped carbon nanocone structures from Transmission Electronic Microscopy (TEM) images is described. The technique enables robust segmentation of the structures by exploiting a simplified Generalized Hough Transform (HT)-based processing. Exploitation of parallelism on commodity hardware is also explored for efficient processing. Effectiveness of the technique is evaluated through experiments on synthetic, simulated, and real images of carbon nanocone structures.