Assisting pictogram selection with categorized semantics

Heeryon Cho*, Toru Ishida, Satoshi Oyama, Rieko Inaba, Toshiyuki Takasaki

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

6 Citations (Scopus)


Since participants at both end of the communication channel must share common pictogram interpretation to communicate, the pictogram selection task must consider both participants' pictogram interpretations. Pictogram interpretation, however, can be ambiguous. To assist the selection of pictograms more likely to be interpreted as intended, we propose a categorical semantic relevance measure which calculates how relevant a pictogram is to a given interpretation in terms of a given category. The proposed measure defines similarity measurement and probability of interpretation words using pictogram interpretations and frequencies gathered from a web survey. Moreover, the proposed measure is applied to categorized pictogram interpretations to enhance pictogram retrieval performance. Five pictogram categories used for categorizing pictogram interpretations are defined based on the five first-level classifications defined in the Concept Dictionary of the EDR Electronic Dictionary. Retrieval performances among not-categorized interpretations, categorized interpretations, and categorized and weighted interpretations using semantic relevance measure were compared, and the categorized semantic relevance approaches showed more stable performances than the not-categorized approach.

Original languageEnglish
Pages (from-to)2638-2646
Number of pages9
JournalIEICE Transactions on Information and Systems
Issue number11
Publication statusPublished - 2008 Nov
Externally publishedYes


  • Categorization
  • EDR
  • Pictogram
  • Semantic relevance

ASJC Scopus subject areas

  • Software
  • Hardware and Architecture
  • Computer Vision and Pattern Recognition
  • Electrical and Electronic Engineering
  • Artificial Intelligence


Dive into the research topics of 'Assisting pictogram selection with categorized semantics'. Together they form a unique fingerprint.

Cite this