Information retrieval through the fusion of multi-sensors data over multi-resolutions

Every sensor is designed to meet a specific objective. To meet the different objectives varieties of sensors are designed. Sometimes the derived information from these specific sensors individually is not enough to meet the objective of current application. There is a need to improve the accuracy of the information with confidence over period of time. In such circumstances, fusion of two or more sensors' data improves the accuracy and confidence in information extraction. In this paper, we propose novel fusion approaches developed by exploring the benefits of the spatial, spectral, temporal and radiometric resolutions of space borne sensors, and multi-modalities of bio sensors. Spatial-spectral fusion method is developed by using Fast Discrete Curvelet Transforms (FDCT). It improves the spatial resolution of a coarser multispectral image. Spatial-radiometric fusion method is developed by using Wavelet Transforms. It improves the radiometric resolution of the high spatial multispectral image. Spatial-temporal fusion method is developed through temporal high pass modulation. It improves the temporal resolution of the high spatial multispectral image. In medical imaging modalities, MRI has an advantage of high spatial resolution compared to the CT, PET, and SPECT imaging modalities. But each one of them has a specific advantage in diagnosing the disease. In this paper, we presented MRI-CT, MRI-PET and MRI-SPECT image fusion approaches by using Wavelet Transforms. Experimental results of MRI-CT image fusion demonstrated which can obtain the benefits of multi-modalities fusion in information retrieval. The tomographic sections of fused images are used for constructing 3D model to derive the complete information. Experimental results of these methods are discussed for the potential opportunities to improve accuracy and confidence in precise information retrieval.