A Method for Dataset Labeling for Activity Recognition in Videos

Eduardo Castro-Giron, Pedro Shiguihara

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

Abstract

Social Sciences and Demographics are in search of multiples ways to analyze person actions besides global factors like the environment. This paper aims to propose a protocol to construct a dataset from videos, then applying it to build a dataset of Peruvian News. Finally, we evaluate it using different Machine Learning (ML) methods, achieving an accuracy ranging 85% and 89%.

Original languageEnglish
Title of host publicationProceedings of the 2021 IEEE 28th International Conference on Electronics, Electrical Engineering and Computing, INTERCON 2021
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781665412216
DOIs
StatePublished - 5 Aug 2021
Externally publishedYes
Event28th IEEE International Conference on Electronics, Electrical Engineering and Computing, INTERCON 2021 - Virtual, Lima, Peru
Duration: 5 Aug 20217 Aug 2021

Publication series

NameProceedings of the 2021 IEEE 28th International Conference on Electronics, Electrical Engineering and Computing, INTERCON 2021

Conference

Conference28th IEEE International Conference on Electronics, Electrical Engineering and Computing, INTERCON 2021
Country/TerritoryPeru
CityVirtual, Lima
Period5/08/217/08/21

Keywords

  • convolutional neural networks
  • labeling
  • video labeling

Fingerprint

Dive into the research topics of 'A Method for Dataset Labeling for Activity Recognition in Videos'. Together they form a unique fingerprint.

Cite this