An Experimentation with Improved Target Vectors for MLP in Classifying Degraded Patterns

Abstract

-In this paper, we adopt unconventional target vectors to improve the performance of pattern classification systems using neural network techniques based on MLP. Instead of conventional target vectors, the new target vectors are bipolar, orthogonal, and highly dimensional. Since they are orthogonal with bipolar representation, we can take advantage of increasing on the Euclidean distance for these vectors when their number (n in a Euclidean space R n ) of components increases. We define non-orthogonal bipolar vectors considered as conventional target vectors for comparison purposes. Those non-orthogonal bipolar vectors provide a fair reference to ensure the effectiveness of the adopted unconventional target vectors and in justifying the credibility and validity of experimental results. The conventional and unconventional target vectors are used in the experiments for training MLP models by backpropagation algorithm to classify patterns extracted from actual degraded images. Comparison of experimental results lead to conclusions that classification performances of MLP models considerably improved with the adopted unconventional target vectors in classifying degraded patterns.

Improving Iris Recognition through New Target Vectors in MLP Artificial Neural Networks

“…The work applied to the recognition of handwritten digits [9] has presented a variation on the recognition rate by 2% when OBVs are used as target vectors for MLP learning. The OBV experimental results for application to the digits extracted from license plate degraded images [10] have presented an increase of 8% on the MLP performance. All these results have strengthened the viability of using OBVs as target vectors for MLP in pattern recognition.…”

“…Preliminary experimental results related to this proposal have been presented by previous works [7][8][9][10] and they have shown the effectiveness for the MLP performance improvement. This paper aims to show experimental results of this new approach applied to iris image recognition.…”