Automatic correction of annotation boundaries in activity datasets by class separation maximization

Reuben Kirkham, Sebastian Mellor, Aftab Khan, Daniel Roggen, Sourav Bhattacharya, Thomas Plötz, Nils Hammerla

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

6 Citations (Scopus)

Abstract

It is challenging to precisely identify the boundary of activities in order to annotate the activity datasets required to train activity recognition systems. This is the case for experts, as well as non-experts who may be recruited for crowd-sourcing paradigms to reduce the annotation effort or speed up the process by distributing the task over multiple annotators. We present a method to automatically adjust annotation boundaries, presuming a correct annotation label, but imprecise boundaries, otherwise known as \label jitter". The approach maximizes the Fukunaga Class-Separability, applied to time series. Evaluations on a standard benchmark dataset showed statistically significant improvements from the initial jittery annotations.

Original languageEnglish
Title of host publicationProceedings of the 2013 ACM conference on Pervasive and ubiquitous computing adjunct publication
Place of PublicationNew York NY USA
PublisherAssociation for Computing Machinery (ACM)
Pages673-678
Number of pages6
ISBN (Print)9781450322157
DOIs
Publication statusPublished - 2013
Externally publishedYes
EventACM International Joint Conference on Pervasive and Ubiquitous Computing 2013 - Zurich, Switzerland
Duration: 8 Sep 201312 Sep 2013

Conference

ConferenceACM International Joint Conference on Pervasive and Ubiquitous Computing 2013
Abbreviated titleUbiComp 2013
CountrySwitzerland
CityZurich
Period8/09/1312/09/13

Keywords

  • Annotation errors
  • Class separability
  • Crowdsourcing
  • Human activity recognition

Cite this