Using Computer Vision and Deep Learning for Nanoparticle Recognition on Scanning Probe Microscopy Images: Modified U-net Approach

Mikhail F. Liz, Anna V. Nartova, Andrey V. Matveev, Aleksey G. Okunev

Research output: Chapter in Book/Report/Conference proceedingConference contributionResearchpeer-review

Abstract

Particles characterization is a significant part of numerous studies in material sciences and engineering technologies. Microscopy images of materials containing particles are usually analyzed by operator with manual counting and measuring of particle sizing by a software ruler. Traditional automated image analyzing methods such as edge detection, segmentation, etc. are not universal, giving poor results on noisy pictures and need empirical fitted parameters. To realize automatic method of particles recognition on scanning tunneling microscopy (STM) data we used U-net and modified U-net neural networks, which was trained on ten STM images contained 1918 particles. Verification on 3 pictures with 695 particles showed mAP=0.12 for modified U-net neural network.

Original languageEnglish
Title of host publicationProceedings - 2020 Science and Artificial Intelligence Conference, S.A.I.ence 2020
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages13-16
Number of pages4
ISBN (Electronic)9780738131115
DOIs
Publication statusPublished - 14 Nov 2020
Event2020 Science and Artificial Intelligence Conference, S.A.I.ence 2020 - Virtual, Novosibirsk, Russian Federation
Duration: 14 Nov 202015 Nov 2020

Publication series

NameProceedings - 2020 Science and Artificial Intelligence Conference, S.A.I.ence 2020

Conference

Conference2020 Science and Artificial Intelligence Conference, S.A.I.ence 2020
CountryRussian Federation
CityVirtual, Novosibirsk
Period14.11.202015.11.2020

Keywords

  • deep neural networks
  • particles recognition
  • scanning probe microscopy

Fingerprint

Dive into the research topics of 'Using Computer Vision and Deep Learning for Nanoparticle Recognition on Scanning Probe Microscopy Images: Modified U-net Approach'. Together they form a unique fingerprint.

Cite this