|
| << vorige | volgende >> |
Tijdschrift beschrijving
Alle jaargangen van het bijbehorende tijdschrift
Alle afleveringen van het bijbehorende jaargang
Alle artikelen van de bijbehorende aflevering
Details van artikel 4 van 13 gevonden artikelen
| Titel: |
Combining Neural Networks For Skin Detection |
| Auteur: |
Chelsia Amy Doukim Jamal Ahmad Dargham Ali Chekima Sigeru Omatu |
| Verschenen in: |
Signal & image processing |
| Paginering: | Jaargang 1 (2011) nr. 2 pagina's 1-11 |
| Jaar: |
2011 |
| Inhoud: |
Two types of combining strategies were evaluated namely combining skin features and combining skinclassifiers. Several combining rules were applied where the outputs of the skin classifiers are combinedusing binary operators such as the AND and the OR operators, “Voting”, “Sum of Weights” and a newneural network. Three chrominance components from the YCbCr colour space that gave the highest correctdetection on their single feature MLP were selected as the combining parameters. A major issue indesigning a MLP neural network is to determine the optimal number of hidden units given a set of trainingpatterns. Therefore, a “coarse to fine search” method to find the number of neurons in the hidden layer isproposed. The strategy of combining Cb/Cr and Cr features improved the correct detection by 3.01%compared to the best single feature MLP given by Cb-Cr. The strategy of combining the outputs of threeskin classifiers using the “Sum of Weights” rule further improved the correct detection by 4.38% comparedto the best single feature MLP. |
| Uitgever: |
Academy & Industry Research Collaboration Center (AIRCC) (provided by DOAJ) |
| Bronbestand: |
Elektronische Wetenschappelijke Tijdschriften |
Details van artikel 4 van 13 gevonden artikelen
| << vorige | volgende >> |