Segmentation and Morphology Computation of a Spiky Nanoparticle Using the Hourglass Neural Network

Abstract

Morphological measurements of nanoparticles in electronmicroscopyimages are tedious, laborious, and often succumb to human errors.Deep learning methods in artificial intelligence (AI) paved the wayfor automated image understanding. This work proposes a deep neuralnetwork (DNN) for the automated segmentation of a Au spiky nanoparticle(SNP) in electron microscopic images, and the network is trained witha spike-focused loss function. The segmented images are used for thegrowth measurement of the Au SNP. The auxiliary loss function capturesthe spikes of the nanoparticle, which prioritizes the detection ofspikes in the border regions. The growth of the particles measuredby the proposed DNN is as good as the measurement in manually segmentedimages of the particles. The proposed DNN composition with the trainingmethodology meticulously segments the particle and consequently providesaccurate morphological analysis. Furthermore, the proposed networkis tested on an embedded system for integration with the microscopehardware for real-time morphological analysis.