A Hierarchical Task Assignment for Manual Image Labeling

Chia Ming Chang, Siddharth Deepak Mishra, Takeo Igarashi

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

Abstract

Manual image labeling (selecting an appropriate 'category' for an image) is very tedious and time consuming especially when selecting labels from a large number of categories. In this study, we propose a hierarchical assignment of labeling tasks where the labelers recursively classify images in a category group into sub category groups, working on a single level at a time. This significantly makes each labeler's task easier, reducing the number of choices from 1,000 to 27 on average. In the user study, we compared our hierarchical assignment to a normal (non-hierarchical) assignment for a labeling task. The results show that the hierarchical assignment requires less total time to complete the labeling task. In addition, the learning effect in the labeling process is more profound in the hierarchical assignment.

Original languageEnglish
Title of host publicationProceedings - 2019 IEEE Symposium on Visual Languages and Human-Centric Computing, VL/HCC 2019
EditorsJustin Smith, Christopher A. Bogart, Judith Good, Scott D. Fleming
PublisherIEEE Computer Society
Pages139-143
Number of pages5
ISBN (Electronic)9781728108100
DOIs
StatePublished - Oct 2019
Event2019 IEEE Symposium on Visual Languages and Human-Centric Computing, VL/HCC 2019 - Memphis, United States
Duration: 14 Oct 201918 Oct 2019

Publication series

NameProceedings of IEEE Symposium on Visual Languages and Human-Centric Computing, VL/HCC
Volume2019-October
ISSN (Print)1943-6092
ISSN (Electronic)1943-6106

Conference

Conference2019 IEEE Symposium on Visual Languages and Human-Centric Computing, VL/HCC 2019
CountryUnited States
CityMemphis
Period14/10/1918/10/19

Keywords

  • hierarchical assignment
  • human-computer interaction
  • image labeling
  • machine learning

Fingerprint Dive into the research topics of 'A Hierarchical Task Assignment for Manual Image Labeling'. Together they form a unique fingerprint.

Cite this