LabMicrobe

LabMicrobe overview
Image analysis
Neural network classifier
Dividing cells
Results
Download LabMicrobe

The neural network classifier

LabMicrobe includes a neural network, which enables it to classify organisms using criteria defined by the user. Below an example of neural network classification of marine bacteria is shown.

Training of the neural network classifier is done by simply pointing and clicking on a nummer of organisms from each class.

Example of a training set used for the neural network classifier. Images were taken of filtered and stained sea water samples from the Baltic Sea. The objects represent the main classes of bacteria in this sea area (R1 - 3: rods, V: vibrio, X: rejected objects). From Blackburn et at, 1998.

The contours of objects are defined using Fourier coeffiecients. The figure shows the levels of details achieved by using different numbers of Fourier coefficients (solid lines). From Blackburn et at, 1998.
Bioras, Munkegårdsvej 6B, DK-3490 Kvistgård, Denmark, tlf. +45 46938088, mobile +45 22678812, e-mail: