Fig. 1. The relationship between entropy and mutual information for two variables The above quantities are schematically demonstrated by the Venn diagram of Figure - Fig. 3. Source images (a) Natural color composite of the first three bands (b) the near infrared band. Results for (c) averaging (d) PCA (e) DWT and (f) morphological fusion methods Fig. 4. Source images (a) Natural color composite of the first three bands (b) the near infrared band. Results for (c) averaging (d) PCA (e) DWT and (f) morphological fusion Table 1. Objective performance evaluation using IFPM and MI for the two datasets 5. Color image fusion Fig. 5. Color image fusion process The objective measures discussed so far address the problem of grayscale image fusion that is fusion methods that result in grayscale representations. However, these measures cannot be trivially extended into color image fusion techniques. image of a scene representing a sandy path, trees and fences and a midwave infrared (3- 5pm) image in which a person is standing behind the trees and close to the fence. Dataset 3 - The third data set originates from the field of remote sensing. It consists of multispectral data acquired from the ENVISAT satellite and specifically MERIS sensor. MERIS (MEdium Resolution Imaging Spectrometer Instrument) measures the solar radiation reflected by the Earth at a ground spatial resolution of 300m, in 15 spectral bands, programmable in width and position, in the visible and near infra-red region of the electromagnetic spectrum. The geographical area is in the South part of Greece and covers both sea and land. The results of the previously described fusion methods can be found in Figure 7. The proposed CIFM vector measure is calculated for the above described color image fusion methods. The two components of the measure namely, /FPM and HD are calculated and the results are summarized in Table 2 for the medical data. In the case of the medical data set Method 2 outperforms other methods for both IFPM and HD vector components whilst Table 2. Objective performance evaluation using IFPM and HD for the different datasets ce * all The two components of the CIFM measure can also be used for a graphical representation in order to evaluate image fusion methods. The CIFM vector components, namely JFPM and HD provide an orthogonal base for a two dimensional vector space (I[FPM, HD) where each fusion method is regarded as a single point. Each vector component can also be used independently in certain applications. For example, IFPM could be employed if the amount of information transferred from the source images to the final image is important since further digital processing will be employed. On the other hand if the fused image will be used by visual experts then special attention should be given to the color distribution and thus HD provide a useful tool. The results of Table 2 are depicted in a _ graphical representation in Figure 8. The same 7 4 @ . «il a a? ow a a oo “ ‘ae! ae The vector components of the CIFM measure are also calculated for the case of the third dataset from the field of remote sensing. Method 2 achieves superior performance in both IFPM and HD components for the case of the third dataset. The fusion method that is based on the wavelet approach is having a comparable performance especially in the color distribution expressed by HD measure. Method 4 achieves rather good but not optimal results in any measure whilst Method 1 fails to provide an efficient performance both in the information transfer but also in the color distribution. These results are compliant with findings reported for the cases of grayscale image fusion aes aa Image fusion technology has successfully contributed to various fields such as medical diagnosis and navigation, surveillance systems, remote sensing, digital cameras, military applications, computer vision, etc. Image fusion aims to generate a fused single image which contains more precise reliable visualization of the objects than any source image of them. This book presents various recent advances in research and development in the field of image fusion. It has been created through the diligence and creativity of some of the most accomplished experts in various fields. Image fusion technology has successfully contributed to various fields such as medical diagnosis and How to reference
