In this paper, the performance of the commonly used neural-network-based classifiers is investigated on solving a classification problem which aims to identify the object nature based on surface features of the object.When the surface data is obtained, a proposed feature extraction method is used to extract the surface feature of the object. The extracted features are then used as the inputs for the classifier. This research studies eighteen household objects which are requisite to our daily life. Six commonly used neural-network-based classifiers, namely one-against-all, weighted one-against-all, binary coded, parallel-structured, weighted parallel structured and tree-structured, are investigated. The performance for the six neural-network-based classifiers is evaluated based on recognition accuracy for individual object. Also, two traditional classifiers, namely k-nearest neighbor classifier and naive Bayes classifier, are employed for the comparison purposes. To evaluate robustness property of the classifiers, the original clean data is contaminated with Gaussian white noise. Experimental results show that the parallel-structured, tree-structured and the naive Bayes classifiers outperform the others under the noise-free data. The tree-structured classifier demonstrates the best robustness property under the noisy data.

This paper presents a novel neural network having variable weights, which is able to improve its learning and generalisation capabilities, to deal with classification problems. The variable weight neural network (VWNN) allows its weights to be changed in operation according to the characteristic of the network inputs so that it can adapt to different characteristics of input data resulting in better performance compared with ordinary neural networks with fixed weights. The effectiveness of the VWNN is tested with the consideration of two real-life applications. The first application is on the classification of materials using the data collected by a robot finger with tactile sensors sliding along the surface of a given material. The second application considers the classification of seizure phases of epilepsy (seizure-free, pre-seizure and seizure phases) using real clinical data. Comparisons are performed with some traditional classification methods including neural network, k-nearest neighbours and naive Bayes classification techniques. It is shown that the VWNN classifier outperforms the traditional methods in terms of classification accuracy and robustness property when the input data is contaminated with noise.

No abstract