Scanning electron microscopy image representativeness: morphological data on nanoparticles.

Area of Science:

• Materials Science
• Nanotechnology
• Computer Vision

Background:

• Nanomaterial characterization requires analyzing diverse particle shapes and sizes.
• Scanning electron microscopy (SEM) images often capture only a fraction of a sample's morphological complexity.
• Quantitative analysis of SEM images is crucial for deriving numerical particle descriptors.

Purpose of the Study:

• To develop a framework for extracting morphological information from SEM images using computer vision algorithms.
• To establish protocols for selecting representative SEM images and determining the minimum set for accurate morphological feature analysis.
• To provide a quantitative method for assessing nanomaterial morphology.

Main Methods:

• Utilized computer vision algorithms to analyze SEM digital images.
• Developed a framework for converting image content into numerical particle descriptors.
• Explored image representativeness and defined protocols for image selection and dataset optimization.

Main Results:

• Successfully extracted morphological information from SEM images.
• Quantified nanoparticle characteristics using numerical descriptors.
• Demonstrated the methodology's practical application on tricalcium phosphate and calcium hydroxyphosphate.

Conclusions:

• The presented framework enables robust quantitative morphological analysis of nanomaterials from SEM images.
• Established protocols ensure image representativeness and optimize data collection for accurate nanomaterial characterization.
• The methodology is applicable to various nanomaterials, including those with biomedical applications.