Automated Extraction of Cellular Features for a Potentially Robust Classification Scheme
Afzal A. Godil, Shwetadwip Chowdhury
The advent of new imaging technologies has allowed faster and higher-resolution cellular image acquisition by light microscopy. This has directly aided clinical research by allowing doctors and other researchers to better visualize cells and cellular components, which improves their ability to provide accurate prognosis, diagnosis, and treatment of a variety of diseases. However, a major bottleneck now is the manual evaluation of such images. Especially with the sheer amount of data being generated from these images, manual evaluation and analysis is becoming increasingly cumbersome. To address this concern, there has been extensive work on researching methods to automatically extract cellular and subcellular features that give morphological and functional insight into the cell. Some examples include extracting size/shape of cell boundaries, density of mitochondria, expression level of proteins, nuclear/cell size ratio, etc. Here, we specifically focus on techniques to robustly extract information about the cellular proteins - actin, myosin, and phosphotyrosine - from images of cells specifically labeled with fluorescent moieties localized to the specific proteins. These proteins are known to be important components in the signaling pathways regulating cellular proliferation. Being able to quickly and automatically extract information about such proteins from raw images will greatly aid biological research, especially on topics related to understanding the role of the microenvironment in disease. Such extracted features can then be used for cellular classification. Future work can incorporate our feature extraction capabilities to classify a cell into categories with known feature parameters.
and , S.
Automated Extraction of Cellular Features for a Potentially Robust Classification Scheme, NIST Interagency/Internal Report (NISTIR), National Institute of Standards and Technology, Gaithersburg, MD, [online], https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=906886
(Accessed December 10, 2023)