Premium
Experimental Computational Evaluation of Biological Immunofluorescence Data
Author(s) -
Hietpas Taylor
Publication year - 2019
Publication title -
the faseb journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.709
H-Index - 277
eISSN - 1530-6860
pISSN - 0892-6638
DOI - 10.1096/fasebj.2019.33.1_supplement.642.6
Subject(s) - computer science , non negative matrix factorization , artificial intelligence , segmentation , image processing , cluster analysis , pattern recognition (psychology) , computer vision , image (mathematics) , matrix decomposition , physics , eigenvalues and eigenvectors , quantum mechanics
Optical spectroscopy and fluorescence have become important techniques for elucidating molecular biological function. Two‐photon and confocal microscopy have particularly enabled increased potential for data acquisition and cellular resolution. With this increased potential has developed a need for efficient and reliable methods for evaluating these large data sets. Historically, researchers have relied upon fundamental image processing techniques and manually implemented image modification to obtain interpretations of their data. While these methods have proven successful, increased computation power and well‐established statistical segmentation protocols have enabled a renaissance of sorts in biological image analysis. In our lab, lymphatic and macrophage immunofluorescence data or endogenous fluorescence from transgenic reporters is most commonly acquired via either confocal or two‐photon microscopy of whole mount corneal and skin tissues. Manual analysis of these large and complex data sets is time consuming and technically challenging. This project has allowed us to explore more sophisticated methods of image processing and analysis. First, K‐means clustering algorithms were implemented to evaluate efficacy of unsupervised partitioning and segmentation of these image data sets. While K‐Means has proven effective at segmentation in many disciplines, the task of segmenting noise and autofluorescence has in these types of data has proven subtler than expected. As such, further methodologies are being investigated including Nonnegative Matrix Factorization, Spectral Phasor Analysis. This abstract is from the Experimental Biology 2019 Meeting. There is no full text article associated with this abstract published in The FASEB Journal .