Ultrasound detection of scatterer concentration by weighted entropy

Po Hsiang Tsui*

*Corresponding author for this work

Research output: Contribution to journalJournal Article peer-review

41 Scopus citations

Abstract

Ultrasound backscattering signals depend on the microstructures of tissues. Some studies have applied Shannon entropy to analyze the uncertainty of raw radiofrequency (RF) data. However, we found that the sensitivity of entropy in detecting various scatterer concentrations is limited; thus, we propose a weighted entropy as a new information entropy-based approach to enhance the performance of scatterer characterization. A standard simulation model of ultrasound backscattering was used to generate backscattered RF signals with different number densities of scatterers. The RF signals were used to estimate the weighted entropy according to the proposed algorithmic scheme. The weighted entropy increased from 0.08 to 0.23 (representing a dynamic range of 0.15) when the number density of scatterers increased from 2 to 32 scatterers/mm2. In the same range of scatterer concentration, the conventional entropy increased from 0.16 to 0.19 (a dynamic range of 0.03). The results indicated that the weighted entropy enables achieving a more sensitive detection of the variation of scatterer concentrations by ultrasound.

Original languageEnglish
Pages (from-to)6598-6616
Number of pages19
JournalEntropy
Volume17
Issue number10
DOIs
StatePublished - 2015

Bibliographical note

Publisher Copyright:
© 2015 by the authors.

Keywords

  • Information entropy
  • Ultrasound
  • Weighted entropy

Fingerprint

Dive into the research topics of 'Ultrasound detection of scatterer concentration by weighted entropy'. Together they form a unique fingerprint.

Cite this