Bacteria classification using neural network

Manual bacteria classification is a tedious work which often needs abundant correlative data and also takes a great deal of time and energy. Combining pattern recognition and new neural network, we propose an approach of bacteria classification based on morphometrics using artificial neural network. The neural network is applied to extract the feature. The entropy sequence is taken as the feature vector. Then a simple classifier is also designed with Euclid distance. The use of relative distance instead of absolute distance improves greatly the accuracy of classification. A mass of experiments are carried out to verify the validity of the proposed method. The results prove that the method is feasible and efficient. This method is also suitable for studying immobilized cell.