Use this URL to cite or link to this record in EThOS:
Title: Signal processing algorithms for enhanced image fusion performance and assessment
Author: Omar, Zaid Bin
ISNI:       0000 0004 2732 0784
Awarding Body: Imperial College London
Current Institution: Imperial College London
Date of Award: 2012
Availability of Full Text:
Access from EThOS:
Access from Institution:
The dissertation presents several signal processing algorithms for image fusion in noisy multimodal conditions. It introduces a novel image fusion method which performs well for image sets heavily corrupted by noise. As opposed to current image fusion schemes, the method has no requirements for a priori knowledge of the noise component. The image is decomposed with Chebyshev polynomials (CP) being used as basis functions to perform fusion at feature level. The properties of CP, namely fast convergence and smooth approximation, renders it ideal for heuristic and indiscriminate denoising fusion tasks. Quantitative evaluation using objective fusion assessment methods show favourable performance of the proposed scheme compared to previous efforts on image fusion, notably in heavily corrupted images. The approach is further improved by incorporating the advantages of CP with a state-of-the-art fusion technique named independent component analysis (ICA), for joint-fusion processing based on region saliency. Whilst CP fusion is robust under severe noise conditions, it is prone to eliminating high frequency information of the images involved, thereby limiting image sharpness. Fusion using ICA, on the other hand, performs well in transferring edges and other salient features of the input images into the composite output. The combination of both methods, coupled with several mathematical morphological operations in an algorithm fusion framework, is considered a viable solution. Again, according to the quantitative metrics the results of our proposed approach are very encouraging as far as joint fusion and denoising are concerned. Another focus of this dissertation is on a novel metric for image fusion evaluation that is based on texture. The conservation of background textural details is considered important in many fusion applications as they help define the image depth and structure, which may prove crucial in many surveillance and remote sensing applications. Our work aims to evaluate the performance of image fusion algorithms based on their ability to retain textural details from the fusion process. This is done by utilising the gray-level co-occurrence matrix (GLCM) model to extract second-order statistical features for the derivation of an image textural measure, which is then used to replace the edge-based calculations in an objective-based fusion metric. Performance evaluation on established fusion methods verifies that the proposed metric is viable, especially for multimodal scenarios.
Supervisor: Stathaki, Tania Sponsor: Kementerian Pengajian Tinggi, Malaysia ; Universiti Teknologi Malaysia
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral