LAUSR

The main issue with the multi-focus images lies in obtaining the relative information about the identification of objects in the individual images with less resolution. Hence the image fusion methods have attracted attention to obtain high resolute image with a pair of multi-focus images. The fusion methods have also raised its attention to apply for biomedical images obtained with different modalities. An attempt has been made in the present work to develop an image fusion methodology designing on discrete cosine transform using frequency partitioning (DCT-FP). For the feature extraction and for better morphological details, the paper discussed about the modified principal component analysis (M-PCA) algorithm. Six sets of medical images obtained with different modalities have been introduced to the six different image fusion algorithms including the proposed method. Various statistical metrics were evaluated for each image fusion method. The careful comparison of the visual and objective metrics reveals that the proposed method shows best performance with not only having visual quality and also confirmed based on the variation of the statistical metrics.