Adaptive principal component analysis fusion schemes for multifocus and different optic condition images

摘要

Fusion of multifocus images and different optic condition (DOC) images deliver a composite image with better information content in their field of application. This paper investigates and analyses the performance of principal component averaging (PCAv) schemes for the fusion of multifocus images and DOC images. Covariance analysis based on localised regions in spatial and transform domain leads to adaptive principal component analysis (PCA). PCAv fusion schemes evaluate linear weights for fusion rule based on adaptive covariance analysis of the source images. Fusion outcomes resulting from this concept prove to be better than conventional PCA and other fusion schemes taken for comparison. Metrics, such as average figure of merit, average quality index and fusion factor, are evaluated to prove the effectiveness of the PCAv fusion methods.