Detection of molecular particles in live cells via machine learning

Shan Jiang, Xiaobo Zhou, Tom Kirchhausen, Stephen T.C. Wong

Research output: Contribution to journalArticlepeer-review

40 Scopus citations


Clathrin-coated pits play an important role in removing proteins and lipids from the plasma membrane and transporting them to the endosomal compartment. It is, however, still unclear whether there exist "hot spots" for the formation of Clathrin-coated pits or the pits and arrays formed randomly on the plasma membrane. To answer this question, first of all, many hundreds of individual pits need to be detected accurately and separated in live-cell microscope movies to capture and monitor how pits and vesicles were formed. Because of the noisy background and the low contrast of the live-cell movies, the existing image analysis methods, such as single threshold, edge detection, and morphological operation, cannot be used. Thus, this paper proposes a machine learning method, which is based on Haar features, to detect the particle's position. Results show that this method can successfully detect most of particles in the image. In order to get the accurate boundaries of these particles, several post-processing methods are applied and signal-to-noise ratio analysis is also performed to rule out the weak spots.

Original languageEnglish (US)
Pages (from-to)563-575
Number of pages13
JournalCytometry Part A
Issue number8
StatePublished - Aug 2007


  • Haar features
  • Machine learning
  • Particle detection
  • Signal-to-noise ratio

ASJC Scopus subject areas

  • Hematology
  • Cell Biology
  • Pathology and Forensic Medicine
  • Biophysics
  • Endocrinology


Dive into the research topics of 'Detection of molecular particles in live cells via machine learning'. Together they form a unique fingerprint.

Cite this