scholarly journals On compact solution vectors in Kronecker-based Markovian analysis

2017 ◽  
Vol 115 ◽  
pp. 132-149 ◽  
Author(s):  
P. Buchholz ◽  
T. Dayar ◽  
J. Kriege ◽  
M.C. Orhan
Keyword(s):  
1997 ◽  
Vol 56 (1) ◽  
pp. 322-330 ◽  
Author(s):  
J. R. Brinati ◽  
S. S. Mizrahi ◽  
G. A. Prataviera
Keyword(s):  

2003 ◽  
Vol 51 (3) ◽  
pp. 303-309 ◽  
Author(s):  
Russell G. Postier ◽  
Megan R. Lerner ◽  
Stan A. Lightfoot ◽  
Rick Vannarath ◽  
Mary M. Lane ◽  
...  

Computer-assisted analysis of DNA ploidy and nuclear morphology were used to elucidate changes in the cell nucleus that occur during the development of experimental pancreatic cancer. Ductal pancreatic adenocarcinoma was induced in 49 Syrian hamsters by SC injection of N-nitrosobis (2-oxopropyl) amine; twenty hamsters served as controls. Groups of animals were sacrificed every 4 weeks for 20 weeks and adjacent sections of pancreatic tissue were H&E and Feulgen-stained for light microscopy and computer assisted cytometry. Pancreatic ductal cells were classified as normal, atypical, or malignant; tissue inflammation (pancreatitis) was also noted when present. DNA ploidy and nuclear morphology evaluation (Markovian analysis) identified an atypical cell stage clearly distinguishable from either normal or malignant cells; pancreatitis preceded this atypia. The DNA ploidy histogram of these atypical cells revealed a major diploid peak and a minor aneuploid peak. The receiver operator characteristic curve areas for a logistic regression model of normal vs atypical cells was 0.94 and for atypical vs malignant was 0.98, numbers indicative of near-perfect discrimination among these three cell types. The ability to identify an atypical cell population should be useful in establishing the role of these cells in the progression of human pancreatic adenocarcinoma.


1976 ◽  
Vol 24 (1) ◽  
pp. 138-144 ◽  
Author(s):  
N J Pressman

Markovian analysis is a method to measure optical texture based on gray-level transition probabilities in digitized images. Experiments are described that investigate that classification performance of parameters generated by Markovian analysis. Results using Markov texture parameters show that the selection of a Markov step size strongly affects classification error rates and the number of parameters required to achieve the maximum correct classification rates. Markov texture parameters are shown to achieve high rates of correct classification in discriminating images of normal from abnormal cervical cell nuclei.


Sign in / Sign up

Export Citation Format

Share Document