[-] Show simple item record

dc.contributor.advisorGanesh, Thiagarajan
dc.contributor.authorAnusha, Katika
dc.date.issued2019
dc.date.submitted2019 Spring
dc.descriptionTitle from PDF of title page viewed June 20, 2019
dc.descriptionThesis advisor: Thiagarajan Ganesh
dc.descriptionVita
dc.descriptionIncludes bibliographical references (pages 143-145)
dc.descriptionThesis (M.S.)--School of Computing and Engineering. University of Missouri--Kansas City, 2019
dc.description.abstractQuantification of cells from immunostained images is a vital procedure in biomedical analysis, as it helps in the measurement of proliferation, immunodetection and differentiation of nuclear markers, which in turn play a significant role in the analysis of the cell functioning. Surgical pathology uses the quantified immunostained images as a diagnostic tool to differentiate between benign and tumor cells. However, manual quantification suffers numerous drawbacks, such as the lack of repeatability due to inter- and intra-observer variability, the lack of precision due to manual visual quantification and the larger time consumption for counting. This led to the introduction of the computerized image counting techniques as a measure to overcome these difficulties. In this context, the present study proposes a software assisted GUI imaging technique and attempts to analyze its efficiency in the quantification of cells. The study adopted various analytical process, such as the comparison in the quantification between manual and automated in different stains, colocalization, to identify the number of active cells while images are spatially overlapped, fusion indexing and the comparison of the cell counts in myotubes with a control value. The comparative analysis between the proposed software assisted imaging technique and manual counting, using different stains, such as β-Gal, DAPI and sclerostin with the help of box plot, yielded a strong significant difference in DAPI and sclerostin stains. No statistically significant differences were observed in β-Gal staining. The descriptive analysis in the quantification of overlapped cell using two overlapped images (β-Gal and DAPI) and three overlapped images (β-Gal, DAPI, sclerostin), demonstrated the improvement in the identification of active cells both in the case of two and three overlapped images. Furthermore, the comparative analysis of the fusion index value of Wnt3a images against control, using fusion bin range and fusion area methods, using box plot, revealed significant variation in fusion index value between Wnt3a and control in fusion area, whereas, the fusion bin did not yield any statistically significant outcome which help the study to reach into a conclusive inference. However, despite promising results, there is scope for improvement, which in turn opens the door for the future researchers to extend the study using more efficient automated imaging systems.eng
dc.description.tableofcontentsIntroduction -- Software description -- Description of program modules -- Procedures with images -- Conclusion
dc.format.extentxix, 146 pages
dc.identifier.urihttps://hdl.handle.net/10355/69006
dc.publisherUniversity of Missouri -- Kansas Cityeng
dc.subject.lcshImaging systems in medicine -- Software
dc.subject.otherThesis -- University of Missouri--Kansas City -- Computer science
dc.titleAutomated Software to Count Stains in Immunostaining Applicationseng
dc.typeThesiseng
thesis.degree.disciplineComputer Science (UMKC)
thesis.degree.grantorUniversity of Missouri--Kansas City
thesis.degree.levelMasters
thesis.degree.nameM.S. (Master of Science)


Files in this item

[PDF]

This item appears in the following Collection(s)

[-] Show simple item record