A Neural Network for classification of chambers arrangement in foraminifera

Foraminifera are very important microfossils to determine geological age of marine rocks. Image analysis techniques are used to compute two set of shape features describing the shape of the most common foraminifera shells. A k-nearest neighbor and a multiplayer perceptron classifiers are compared for automated classification of the chambers arrangement. Experimental results show 87.1 and 97.1% of accuracy using, respectively, k-nearest neighbor and multiplayer perceptron.