Feature selection, matching, and evaluation for subcellular structure tracking.

Quan Wen, Jean Gao, Kate Luby-Phelps

Research output: Contribution to journalArticlepeer-review

Abstract

Understanding the motility of subcellular particles like organelles, vesicles, or mRNAs is critical to understand how cells regulate delivery of specific proteins from the site of synthesis to the site of action. The goal of this paper is to present a framework of feature selection, matching, and evaluation for the segmentation and tracking of green fluorescent protein (GFP) labeled subcellular structures. To select stable and distinctive features for small-sized subcellular particles, a grid-based minimum variance (GMV) feature selection method is proposed. To robustly keep tracking of the selected features, we propose a mean minimum to maximum ratio (MMMR) similarity measure for feature matching. In order to quantitatively evaluate the proposed methods, we define two evaluation criteria, feature convergence rate (FCVR) and feature consistence rate (FCSR), which conform with the proximity and similarity properties of Gestalt visual perception theory. Our technique was validated on real confocal video data with comparison to traditional feature selection and matching methods.

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition
  • Signal Processing
  • Biomedical Engineering
  • Health Informatics

Fingerprint

Dive into the research topics of 'Feature selection, matching, and evaluation for subcellular structure tracking.'. Together they form a unique fingerprint.

Cite this