Automatic morphometric analysis of neural cells

• Autorzy: Kim J. , Gillies D.
• Języki publikacji: EN

Abstrakty

EN

Studies of developing neural cells are widely used in fundamental research into the mechanisms of nervous diseases, such as multiple sclerosis. These studies nirmally require researchers to estimate the populations of different classes of cells in microscope images of either tissue or cultures. The estimation process is carried out by human visual inspection, and is time comsuming and subjective in nature. As an attack on this problem, we have been investigating the use of computer vision to classify and count the neural cells automatically. We apply various image processing techniques to reduce the neural cells in an image to a skeleton form. Then various measures such as the fractal dimension and 2nd moment are applied to classifi cells according to thier spatial growth characteristics seen during thier development. The measures are then combined using a bayesian classifier, and a decision is taken as to which class a cell belongs. Out initial studies were aimed at classifying the different development stages of the cells knowns as oligodendrocytes. The results were encouraging with better then 80% agreement between the computer analysis and human judgement.

Słowa kluczowe

EN

Sholl coeficient; fractal dimension; 2 nd moment; profile; byesian classifier; oligodendrocytes

• Wydawca: Institute of Information Technology of the Warsaw University of Life Sciences – SGGW
• Czasopismo: Machine Graphics and Vision
• Rocznik: 1998
• Tom: Vol. 7, No. 4
• Strony: 693--709
• Opis fizyczny: Bibliogr. 19 poz.

Twórcy

autor

Kim J.

autor

Gillies D.

• Department of Computing, Imperial College of LONDON, UK