Automated Quantitative Analysis of Silica Nanorod Dimensions via Watershed Segmentation

Abstract

Quantifying the dimensions of silica nanorods often requires manual analysis of their dimensions, but this method is time consuming and tedious. This work explores the potential for an automated analysis with Matlab to improve the efficiency of this analysis. The program described is a preliminary proof-of-concept version of a nanorod analysis program. Watershed segmentation and bounding boxes are viable tools for the automated quantitative analysis of nanorod dimensions, and the automated process is far quicker than manual analysis. While the automated process shows promise, the program functions best with minimal nanorod overlap and requires more extensive testing to become feasible for widespread use. Improvements to noise reduction and particle shape prediction will expand the scope of images that can be subject to automatic analysis.