Image Fusion: New Era in the Field of Digital and Medical Imaging
In the past decade the principle idea of image fusion is gaining large momentum in the research and development in order to generate a fused image which carries much more comprehendible and accurate information. In the process of image fusion, to compute the fused pixel value either the weighted average of the input pixels is calculated where the weight given the pixel could be adaptively according to the algorithm. Where image fusion has a wide spread application in various fields, it becomes of significant importance in the medical imaging. Medical image fusion aims at improvising the image content obtained from varied imaging sensors like Computed Tomography (CT), Positron Emission Tomography (PET) and Magnetic Resonance Imaging (MRI) and combining them into a more useful source of information. In the past few years computing power has advanced sufficiently to finally enable real-time image fusion systems to become a reality and the field has started to move out of the research laboratories and into real products. Algorithmic techniques for fusing images are now well known and understood, challenges remain with regard to exploiting different sensor modalities, robustness to environmental and operational conditions and cost cutting. The various reputed journals are the premier vehicles for disseminating information on all aspects of research and development in the field of image fusion. The advancement in the signal processing theory has lead to documentation of several important mathematical tools for fusion of images like wavelets, sparse representation methods, spatial domain methods and artificial intelligence. The fused image is a result which contains all the critical information of the source images thereby making them more suitable for general human perception and machine vision tasks. The need for image fusion stems from the inherent limitation of optical depths, design and observational constraints of a single sensor. Multiple images are therefore required in order to suffice for the information left by a single sensor. Various source modalities employed tend to exhibit various characteristic features like colours, spectral bands, texture properties, complementary information in a single band. Multi-focus images are taken repeatedly with various focal lengths to create one sharp image. Same camera at the same position captures images at different times is used to detect changes, to enhance night-time scenery, or to restore an image from blurred and noisy image sequence. Multi-exposure is the superimposition of images captured from various exposures to create a single image used for context enhancement or for increasing dynamic range. This blog is intended to present general wave of introduction to the field of multi-process information fusion and hence to promote synergism amongst the masses of many disciplines to contribute to its growth.