Mohamed Abd Elfattah

CT Liver Segmentation Using Artificial Bee Colony Optimisation

Abstract The automated segmentation of the liver area is an essential phase in liver diagnosis from medical images. In this paper, we propose an artificial bee colony (ABC) optimisation algorithm that is used as a clustering technique to segment the liver in CT images. In our algorithm, ABC calculates the centroids of clusters in the image together with the region corresponding to each cluster. Using mathematical morphological operations, we then remove small and thin regions, which may represents flesh regions around the liver area, sharp edges of organs or small lesions inside the liver. The extracted regions are integrated to give an initial estimate of the liver area. In a final step, this is further enhanced using a region growing approach. In our experiments, we employed a set of 38 images, taken in pre-contrast phase, and the similarity index calculated to judge the performance of our proposed approach. This experimental evaluation confirmed our approach to afford a very good segmentation accuracy of 93.73% on the test dataset.

Nature Inspired Optimization Algorithms for CT Liver Segmentation

Nature inspired optimization algorithms have gained popularity in the last two decades due to their efficiency and flexibility when they applied to solve global optimization problems. These algorithms are inspired from the biological behavior by swarms of birds, fish and bees. In this chapter, we give an overview of some of nature inspired optimization algorithms such as Artificial Bee Colony (ABC), Cuckoo Search (CS), Social Spider Optimization (SSO) and Grey Wolf Optimization (GWO). Also, we present the usage of ABC and GWO algorithms for CT liver segmentation. The experimental results of the two selected algorithms show that the two algorithms are powerful and can obtain good results when applied to segment medical images.

Enhanced Region Growing Segmentation for CT Liver Images

This paper intends to enhance the image for the next usage of region growing technique for segmenting the region of liver away from other organs. The approach depends on a preprocessing phase to enhance the appearance of the boundaries of the liver. This is performed using contrast stretching and some morphological operations to prepare the image for next segmentation phase. The approach starts with combining Otsu’s global thresholding with dilation and erosion to remove image annotation and machine’s bed. The second step of image preparation is to connect ribs, and apply filters to enhance image and deepen liver boundaries. The combined filters are contrast stretching and texture filters. The last step is to use a simple region growing technique, which has low computational cost, but ignored for its low accuracy. The proposed approach is appropriate for many images, where liver could not be separated before, because of the similarity of the intensity with other close organs. A set of 44 images taken in pre-contrast phase, were used to test the approach. Validating the approach has been done using similarity index. The experimental results, show that the overall accuracy offered by the proposed approach results in 91.3 % accuracy.

Wolf Local Thresholding Approach for Liver Image Segmentation in CT Images

This paper enhances the usage of level set method to get a reliable liver image segmentation in CT images. The approach depends on a preprocessing phase to enhance the liver’s edges. This phase is performed in two ways using the morphological operations and wolf local thresholding. The first way starts with applying the morphological operations on the image to clean image annotation and bed lines. Then, it applies contrast stretching and texture filters. The other way applies the wolf local threshold to each point in the image. It uses a window or a mask to calculate the average and standard deviation to apply an iterative threshold. Each way is followed by a step of connecting ribs to separate the flesh and skin from liver’s region. The last step is to use level set method to segment the whole liver. A set of 47 images taken in pre-contrast phase, were used to test the approach. Validating the approach is done using similarity index measure. The obtained experimental results showed that the overall accuracy presented by the proposed approach results in 93.19 % accuracy for using morphological operations, and 93.30 % accuracy for using Wolf local thresholding.

Handwritten Arabic Manuscript Image Binarization Using Sine Cosine Optimization Algorithm

Historic manuscript image binarization is considered an important step due to the different kinds of degradation effects on optical character recognition (OCR) or word spotting systems. Previous methods failed on to find the optimal threshold for binarization. In this paper, we investigate the effects of sine cosine algorithm (SCA) on reducing the compactness K-means Clustering as the objective function. The SCA searches for the optimal clustering of the given handwritten manuscript image into compact clusters under some constraints. The proposed approach is evaluated and assessed on a set of selected handwritten Arabic manuscript images. The Experimental result shows that the proposed approach provides the highest value than the famous binarization methods such as; Otsu’s and Niblack’s in terms of F-measure, Pseudo- F-measure, PSNR, Geometric accuracy and the low value on DRD, NRM, MPM.

Region Growing Segmentation with Iterative K-means for CT Liver Images

In this paper, it is intended to enhance the simple region growing technique (RG) to extract liver from the abdomen away from other organs in CT images. Iterative K-means clustering technique is used as a preprocessing step to pass the image to region growing and watershed segmentation techniques. The usage of K-means and region growing is preferred here for its simplicity and low cost of execution. The proposed approach starts with cleaning the annotation and enhancing the boundaries of the liver. This is performed using texture filter and ribs connection algorithm, followed by iterative K-means. K-means removes the clusters with higher intensity values. Then region growing is used to separate the whole liver. Finally, comes the role of watershed that divides the liver into a number of regions of interest (ROIs). The experimental results show that the overall accuracy offered by the proposed approach, results in 92.38% accuracy.

Artificial Bee Colony Based Segmentation for CT Liver Images

The objective of this paper is to evaluate an approach for CT liver image segmentation, to separate the liver, and segment it into a set of regions of interest (ROIs). The automated segmentation of liver is an essential phase in all liver diagnosis systems for different types of medical images. In this paper, the artificial bee colony optimization algorithm (ABC) aides to segment the whole liver. It is implemented as a clustering technique to achieve this mission. ABC calculates the centroid values of image clusters in CT images. Using the least distance between every pixel value and different centroids will result in a binary image for each cluster. Applying some morphological operations on every binary clustered image can help to remove small and thin objects. These objects represent parts of flesh tissues adjacent to the liver, sharp edges of other organs and tiny lesions spots inside the liver. This is followed by filling the large regions in each cluster binary image. Summation of the clusters’ binary images results in a reasonable image of segmented liver. Then, the segmented image of liver is enhanced using simple region growing technique (RG). Finally, one of ABC algorithm or watershed is applied once to extract the lesioned regions in the liver, which can be used by any classifier to determine the type of lesion. A set of 38 images, taken in pre-contrast phase, was used to segment the liver and test the proposed approach. Testing the results is handled using similarity index to validate the success of the approach. The experimental results showed that the overall accuracy offered by the proposed approach, results in 93.73 % accuracy.

Artificial bee colony optimizer for historical Arabic manuscript images binarization

Historical manuscript image binarization is a very important step towards full word spotting system. In this paper, we present a novel binarization algorithm based on artificial bee colony optimizer. The proposed approach contains two phases. The first phase is stretching the intensity level of the image by contrast stretching filter and removing the noise by image cleaning algorithm, the second phase is determining the number of clusters, number of colony and iterations for starting Artificial Bee Colony (ABC) algorithm. The proposed approach is tested on a set of images collected from the electronic Arabic manuscripts database and compared against three famous binarization methods such as Niblacks, Otsus and Savouls. The Experimental results show that the proposed approach is a promising approach and can obtain the desired results better than the other compared methods.

An intelligent approach for galaxies images classification

This article presents an intelligent automatic approach for galaxies images classification based on Artificial Neural Network (ANN) and moment-based features extraction algorithms. The proposed approach consists of three phases; namely, image denoising, feature extraction, and classification phases. For the denoising phase, noise pixels are removed from input images, then input galaxy image is normalized to a uniform scale and Hu seven invariant moment algorithm is applied to reduce the dimensionality of the feature space during the feature extraction phase. Finally, during the classification phase, Self-Organize Feature Maps (SOFMs) and Time Lag Recurrent Networks (TLRNs) algorithms are utilized for classifying the input galaxies images into one of four obtained source catalogue types. Experimental results showed that SOFMs provided better classification results than having TLRNs applied. It is also concluded that a small set of features is sufficient to classify galaxy images and provide a fast classification.

Automated classification of galaxies using invariant moments

Classification and identification of galaxy shape is an important issue for astronauts since it provides valuable information about the origin and the evolution of the universe. Statistical invariant features that are functions of moments have been used as global features of galaxy images in their pattern recognition. In this paper, an automated training based recognition system that can compute the statistical invariant features for different galaxy shapes is investigated. The proposed algorithm is robust, regardless of orientation, size and position of the galaxy inside the image. Feature vectors are computed via nonlinear moment invariant functions for each galaxy shape. After feature extraction, the recognition performance of classifier in conjunction with these moment---based features is introduced. Computer simulations show that Galaxy images are classified with an accuracy of about 90% compared to the human visual classification system.