Multilayer feed forward neural networks for olive trees identification

Abstract EN

In this paper we study and propose Multilayer Feed-Forward Neural Networks for identification some of different olive varieties.

The identification depends on image features of olive tree leaves.

The Feed forward Neural Networks is used with two hidden layers and one output layer with 12 nodes that correspond to different varieties (classes) of olive leaves.

The proposal technique is tested on an data base of 120 images leaves with 10 images for each variety (class).

The result shows that our technique is promising where the recognition rates 100%, and 91% for the training and testing respectively