Computational methods for automated mitosis detection in histopathology images: A review

Tojo Mathew, Jyoti R. Kini, Jeny Rajan

Research output: Contribution to journalReview articlepeer-review

7 Citations (Scopus)


Mitosis detection is an important step in pathology procedures in the context of cancer diagnosis and prognosis. Prevalent process for this task is by manually observing Hematoxylin and Eosin (H & E) stained histopathology sections on glass slides through a microscope by trained pathologists. This conventional approach is tedious, error-prone, and has shown high inter-observer variability. With the advancement of computational technologies, automating mitosis detection by the use of image processing algorithms has attracted significant research interest. In the past decade, several methods appeared in the literature, addressing this problem and they have shown encouraging incremental progress towards a clinically usable solution. Mitosis count is an important parameter in grading of breast cancer and glioma, unlike other cancer types. Driven by the availability of multiple public datasets and open contests, most of the methods in literature address mitosis detection in breast cancer images. This paper is a comprehensive review of the methods published in the area of automated mitotic cell detection in H & E stained histopathology images of breast cancer in the last 10 years. We also discuss the current trends and future prospects of this clinically relevant task, augmenting humanity's fight against cancer.

Original languageEnglish
Pages (from-to)64-82
Number of pages19
JournalBiocybernetics and Biomedical Engineering
Issue number1
Publication statusPublished - 01-01-2021

All Science Journal Classification (ASJC) codes

  • Biomedical Engineering


Dive into the research topics of 'Computational methods for automated mitosis detection in histopathology images: A review'. Together they form a unique fingerprint.

Cite this