Zaid Omar

Wavelet-based medical image fusion via a non-linear operator

Medical image fusion has been extensively used to aid medical diagnosis by combining images of various modalities such as Computed Tomography (CT) and Magnetic Resonance Image (MRI) into a single output image that contains salient features from both inputs. This paper proposes a novel fusion algorithm through the use of a non-linear fusion operator, based on the low sub-band coefficients of the Discrete Wavelet Transform (DWT). Rather than employing the conventional mean rule for approximation sub-bands, a modified approach is taken by the introduction of a non-linear fusion rule that exploits the multimodal nature of the image inputs by prioritizing the stronger coefficients. Performance evaluation of CT-MRI image fusion datasets based on a range of wavelet filter banks shows that the algorithm boasts improved scores of up to 92% as compared to established methods. Overall, the non-linear fusion rule holds strong potential to help improve image fusion applications in medicine and indeed other fields.

Automated detection of microaneurysm for fundus images

Chronic hyperglycemia of diabetes may lead to failure in various organs, especially the blood vessels, eyes, kidneys, nerves and heart. It happens due to the faults either in insulin secretion, insulin action, or both. As diabetes developed, the vision of patient starts to deteriorate, leading to Diabetic Retinopathy (DR). Microaneurysm (MA) are the earliest sign of DR where it appears in clusters as tiny, dark red spots or tiny hemorrhages-like within the retina light-sensitive area. Thus, the objectives of this study are to develop an automated algorithm to perform early detection of MA presence in fundus images, and to evaluate the performance of the proposed system design by evaluating the accuracy of the segmented MA. The methods involved in the pre-processing stage are the green component extraction and bottom hat filtering with gamma correction. As the characteristics of blood vessel and MA are the same, the extraction of vessels is needed. This is done by applying the Gaussian matched filter. It is then segmented out by using certain threshold value. In template learning, wavelet coefficient is used in separating the pattern and background image by following the Gaussian distribution curve. Texture energy filter is used to extract the true features where MA are identified. As a result, 84.15% of accuracy is obtained.

Efficient Motion Estimation Algorithms for HEVC/H.265 Video Coding

This paper presents two fast motion estimation algorithms based on the structure of the triangle and the pentagon, respectively, for HEVC/H.265 video coding. These new search patterns determine motion vectors faster than the two Tzsearch patterns - diamond and square - that are built into the motion estimation engine of the HEVC. The proposed algorithms are capable of achieving a faster run-time with negligible video quality loss and increase in bit rate. Experimental results show that, at their best, the triangle and pentagon algorithms can offer 63 % and 61.9 % speed-up in run-time respectively compared to the Tzsearch algorithms in HEVC reference software.

Combination of EEMD and Morphological Filtering for Baseline Wander Correction in EMG Signals

This paper aims at proposing an effective method for Baseline Wander removal from the EMG signals. Ensemble Empirical Mode Decomposition (EEMD) Algorithm is first applied to the baseline corrupted EMG signals to decompose them into Intrinsic Mode Functions (IMFs). After this step, morphological filtering employing octagon-shaped structuring element has been applied to filter out each IMF. Finally, the results of the proposed filtering methodology are compared with those of EMD- and EEMD-based filtering methods. Simulation results report that the methodology used in this study has eliminated the baseline wander from EMG signals with minimal distortions.

Aortic annulus sizing using watershed transform and morphological approach for CT images

Aortic valve disease occurs due to calcification deposits on the area of leaflets within the human heart. It is progressive over time where it can affect the mechanism of the heart valve. To avoid the risk of surgery for vulnerable patients especially senior citizens, a new method has been introduced: Transcatheter Aortic Valve Implantation (TAVI), which places a synthetic catheter within the patient’s valve. This entails a procedure of aortic annulus sizing, which requires manual measurement of the scanned images acquired from Computed Tomographic (CT) by experts. The step requires intensive efforts, though human error may still eventually lead to false measurement. In this research, image processing techniques are implemented onto cardiac CT images to achieve an automated and accurate measurement of the heart annulus. The image is first put through pre-processing for noise filtration and image enhancement. Then, a marker image is computed using the combination of opening and closing operations where the...

An alphabetic contour-based descriptor for shape-based image retrieval

Content-based image retrieval methods use color, texture and shape of an object for indexing and retrieval. Among these features, shape is a basic visual feature that holds significant information of the object. In this paper an alphabetic contour-based shape description method is proposed to facilitate shape classification and retrieval. The proposed method breaks down shapes contour into small segments and assigns unique alphabetic symbol for each segment based on its geometrical features. These symbols are used to create feature string which we call it, an alphabet string. The alphabet strings are compared together using dynamic programming during classification. The proposed method was tested on BROWN dataset that consists of occluded, articulated and missing part shapes. Results show the feasibility of the method and highlight its advantages over some state-of-the art methods.

Wavelet-based aortic annulus sizing of echocardiography images

Aortic stenosis (AS) is a condition where the calcification deposit within the heart leaflets narrows the valve and restricts the blood from flowing through it. This disease is progressive over time where it may affect the mechanism of the heart valve. To alleviate this condition without resorting to surgery, which runs the risk of mortality, a new method of treatment has been introduced: Transcatheter Aortic Valve Implantation (TAVI), in which imagery acquired from real-time echocardiogram (Echo) are needed to determine the exact size of aortic annulus. However, Echo data often suffers from speckle noise and low pixel resolution, which may result in incorrect sizing of the annulus. Our study therefore aims to perform an automated detection and measurement of aortic annulus size from Echo imagery. Two stages of algorithm are presented — image denoising and object detection. For the removal of speckle noise, Wavelet thresholding technique is applied. It consists of three sequential steps; applying linear discrete wavelet transform, thresholding wavelet coefficients and performing linear inverse wavelet transform. For the next stage of analysis, several morphological operations are used to perform object detection as well as valve sizing. The results showed that the automated system is able to produce more accurate sizing based on ground truth.

A Hybrid Chebyshev-ICA Image Fusion Method Based on Regional Saliency

An image fusion method that performs robustly for image sets heavily corrupted by noise is presented in this paper. The approach combines the advantages of two state-of-the-art fusion techniques, namely Independent Component Analysis (ICA) and Chebyshev Poly-nomial Analysis (CPA) fusion. Fusion using ICA performs well in transferring the salient features of the input images into the composite output, but its performance deteriorates severely under mild to moderate noise conditions. CPA fusion is robust under severe noise conditions, but eliminates the high frequency information of the images involved. We pro-pose to use ICA fusion within high activity image areas, identified by edges and strong textured surfaces and CPA fusion in low activity areas identified by uniform background regions and weak texture. A binary image map is used for selecting the appropriate method, which is constructed by a standard edge detector followed by morphological operators. The results of the proposed approach are very encouraging as far as joint fusion and denoising is concerned. The works presented may prove beneficial for future image fusion tasks in real world applications such as surveillance, where noise is heavily present.

Multimodal image registration for potential diagnosis and monitoring of morphoea using a hybrid NGC method

In medicine, thermal imaging diagnostic tool is increasingly employed thanks to its low cost, non-harmful and non-invasive nature. However, a thermogram can be quite ambiguous due to its low spatial resolution. This ambiguity can be reduced by incorporating information or data from different imaging sensors. This paper is an extension of a modified Normalized Gradient Correlation (NGC) employed in the initial phase of this study for multimodal image registration to assist in diagnosis and monitoring of linear morphoea. The proposed method is an improved, hybrid version of the modified NGC that incorporates an iterative based normalized cross-correlation coefficient (NCC) method for retrieval of translational differences based on the spatial domain in the initial method. The hybrid NGC method is found to reduce misregistration due to inaccurate retrieval of translational differences suffered by the initial NGC method in this multimodal image registration by up to 77.4% for over-detection error.

Efficient morphological-based edge detection algorithm for computed tomography cardiac images

Medical imaging has been extensively used for disease monitoring, treatment planning, diagnosis and computer aided surgery. Often, the acquired images are raw in nature, thus making them prone to being complex and noisy. A series of preprocessing and information extraction steps are therefore necessary in order for the relevant information to reach the medical practitioner. To this end, image denoising and edge detection play a vital role as a precursor to more advanced techniques in the medical image processing field. In this paper we have proposed an innovative mathematical morphology-based image denoising and edge detection method, for pre-processing of computed tomography (CT) images of the human heart. The morphological edge detection algorithm together with six different shaped structuring elements are implemented to preserve and detect the edges of the CT image while effectively suppressing noise, all at low computational cost. The experimental results affirm our approachs efficiency and capability in denoising and detecting salient edges from corrupted and complex medical images.