Abstract
A texture image classification system is presented based upon the use of two cascaded multi-layer feedforward neural networks. The first network transforms a set of high-dimensional and correlated feature images into another set of dimensionally compressed, uncorrelated principal feature images whilst reducing computation effort and minimizing the information lost. The second accomplishes the task of feature pattern classification by using only those principal features obtained by the former. Such a cascaded use of neural networks significantly simplifies the structure of the classification network and increases the efficiency of the overall classification process. These are demonstrated by the comprehensive results obtained from practical applications of the system.
Original language | English |
---|---|
Pages (from-to) | 675-687 |
Number of pages | 13 |
Journal | Pattern Recognition |
Volume | 27 |
Issue number | 5 |
DOIs | |
Publication status | Published - 1994 |
Keywords
- Classification performance
- Neural networks
- Principal component transformation
- Principal feature extraction
- Texture classification