Muhammad Mohsin Riaz

Dual-tree complex wavelet transform and SVD based medical image resolution enhancement

Abstract Dual-tree complex wavelet transform, non-local means filter and singular value decomposition based medical image resolution enhancement is proposed. Contrast of the input image is enhanced using proposed singular value decomposition and high frequency subbands are obtained using dual-tree complex wavelet transform. The contrast enhanced low resolution image and high frequency subbands are interpolated using Lanczos interpolator. Non-local means filter is used to cater the artifacts produced by dual-tree complex wavelet transform. Interpolated contrast enhanced low resolution image and filtered high frequency subbands are combined using inverse dual-tree complex wavelet transform to obtain contrast enhanced super resolution image. Quantitative and qualitative analysis is used to justify the significance of the proposed technique.

Principle Component Analysis and Fuzzy Logic Based through Wall Image Enhancement

Principle component analysis based through wall image enhancement is proposed which is capable of discriminating target, noise and clutter signals. The overlapping boundaries of clutter, noise and target signals are separated using fuzzy logic. Fuzzy inference engine is used to assign weights to principle components. The proposed scheme works well signiflcantly for extracting multiple targets having difierent range proflles in heavy cluttered through wall images. Simulation results are compared on the basis of mean square error, peak signal to noise ratio and visual inspection.

Through-Wall Image Enhancement Based on Singular Value Decomposition

Singular value decomposition and information theoretic criterion-based image enhancement is proposed for through-wall imaging. The scheme is capable of discriminating target, clutter, and noise subspaces. Information theoretic criterion is used with conventional singular value decomposition to find number of target singular values. Furthermore, wavelet transform-based denoising is performed (to further suppress noise signals) by estimating noise variance. Proposed scheme works also for extracting multiple targets in heavy cluttered through-wall images. Simulation results are compared on the basis of mean square error, peak signal to noise ratio, and visual inspection.

Improved Guided Image Fusion for Magnetic Resonance and Computed Tomography Imaging

Improved guided image fusion for magnetic resonance and computed tomography imaging is proposed. Existing guided filtering scheme uses Gaussian filter and two-level weight maps due to which the scheme has limited performance for images having noise. Different modifications in filter (based on linear minimum mean square error estimator) and weight maps (with different levels) are proposed to overcome these limitations. Simulation results based on visual and quantitative analysis show the significance of proposed scheme.

MRI and PET image fusion using fuzzy logic and image local features.

An image fusion technique for magnetic resonance imaging (MRI) and positron emission tomography (PET) using local features and fuzzy logic is presented. The aim of proposed technique is to maximally combine useful information present in MRI and PET images. Image local features are extracted and combined with fuzzy logic to compute weights for each pixel. Simulation results show that the proposed scheme produces significantly better results compared to state-of-art schemes.

Guided Filter and IHS-Based Pan-Sharpening

A guided filter and intensity-hue-saturation-based pan-sharpening scheme is proposed. The scheme combines the high-resolution unispectral and low-resolution multispectral images considering the intensity levels and the spatial information. Guided filtering is used to further refine the weight maps for each pixel. The simulation results show that the suggested scheme mostly yields superior results compared with the existing schemes.

Image Dehazing Using Quadtree Decomposition and Entropy-Based Contextual Regularization

In this letter, an improved single image dehazing technique based on quadtree decomposition and entropy-based weighted contextual regularization is proposed. The boundary constraints are computed adaptively using statistical properties of image. The proposed technique produces high-quality dehazed image with better colors and minimal blocking artifacts. Simulation results compared visually and quantitatively with state-of-the-art existing schemes show the significance of proposed technique.

Adaptive Compressive Fusion for Visible/IR Sensors

An image fusion scheme is proposed for visible and infrared sensors, which adaptively adjusts the number of compressive measurements depending on the amount of information. Simulation results show that the proposed scheme is a significant improvement compared with existing schemes.

MRI Brain Classification Using Texture Features , Fuzzy Weighting and Support Vector Machine

A technique for magnetic resonance brain image classifl- cation using perceptual texture features, fuzzy weighting and support vector machine is proposed. In contrast to existing literature which generally classifles the magnetic resonance brain images into normal and abnormal classes, classiflcation with in the abnormal brain which is relatively hard and challenging problem is addressed here. Texture features along with invariant moments are extracted and the weights are assigned to each feature to increase classiflcation accuracy. Multi- class support vector machine is used for classiflcation purpose. Results demonstrate that the classiflcation accuracy of the proposed scheme is better than the state of art existing techniques.

SAR image segmentation based on active contours with fuzzy logic

In this paper, a fuzzy weighted active contour model for synthetic aperture radar image segmentation is proposed. A feature-based approach comprising weighted active contours and level set segmentation is used to improve segmentation results. A fuzzy inference engine is used to assign weights to pixels of the level set function based on local entropy and local variance. Experiment results show that the proposed scheme improves the segmentation efficiency and accuracy as compared with the existing region-scalable fitting scheme.