Automated Bright Field Segmentation of Cells and Vacuoles Using Image Processing Technique

Pei Ju Chiang, Shao Ming Wu, Min Jen Tseng, Pin Jie Huang

Research output: Contribution to journalArticlepeer-review

3 Citations (Scopus)

Abstract

Understanding the mechanisms and other variants of programmed cell death will help provide deeper insight into various disease processes. Although complex procedures are required to distinguish each type of cell death, the formation of vacuoles is one of the important features in some process of cell death under different conditions. Thus, monitoring and counting the number of vacuoles and the ratio of cells with vacuoles is a commonly used method to indicate and quantify the efficacy of the therapy. Several studies have shown that image processing can provide a quick, convenient and precise mean of performing cell detection. Hence, this study uses an image processing technique to detect and quantify vacuolated cells without the need for dyes. The system both counts the number of vacuolated cells and determines the ratio of cells with vacuoles. The performance of the proposed image processing system was evaluated using 38 images. It has been shown that a strong correlation exists between the automated counts and the manual counts. Furthermore, the absolute percentage errors between automated counts and manual counts for cell detection and vacuolated cell detection using data pooled from all images are 3.61 and 3.33%, respectively. A user-friendly graphical user interface (GUI) is also developed and freely available for download, providing researchers in biomedicine with a more convenient instrument for vacuolization analysis.

Original languageEnglish
Pages (from-to)1004-1018
Number of pages15
JournalCytometry Part A
Volume93
Issue number10
DOIs
Publication statusPublished - 2018 Oct

All Science Journal Classification (ASJC) codes

  • Pathology and Forensic Medicine
  • Histology
  • Cell Biology

Fingerprint Dive into the research topics of 'Automated Bright Field Segmentation of Cells and Vacuoles Using Image Processing Technique'. Together they form a unique fingerprint.

Cite this