Abder-Rahman Ali

A Novel Evolutionary Swarm Fuzzy Clustering Approach for Hyperspectral Imagery

In land cover assessment, classes often gradually change from one to another. Therefore, it is difficult to allocate sharp boundaries between different classes of interest. To overcome this issue and model such conditions, fuzzy techniques that resemble human reasoning have been proposed as alternatives. Fuzzy C-means is the most common fuzzy clustering technique, but its concept is based on a local search mechanism and its convergence rate is rather slow, especially considering high-dimensional problems (e.g., in processing of hyperspectral images). Here, in order to address those shortcomings of hard approaches, a new approach is proposed, i.e., fuzzy C-means which is optimized by fractional order Darwinian particle swarm optimization. In addition, to speed up the clustering process, the histogram of image intensities is used during the clustering process instead of the raw image data. Furthermore, the proposed clustering approach is combined with support vector machine classification to accurately classify hyperspectral images. The new classification framework is applied on two well-known hyperspectral data sets; Indian Pines and Salinas. Experimental results confirm that the proposed swarm-based clustering approach can group hyperspectral images accurately in a time-efficient manner compared to other existing clustering techniques.

Fuzzy C-Means Based Liver CT Image Segmentation with Optimum Number of Clusters

In this paper, we investigate the effect of using an optimum number of clusters with Fuzzy C-Means clustering, for Liver CT image segmentation. The optimum number of clusters to be used was measured using the average silhouette value. The evaluation was carried out using the Jaccard index, in which we concluded that using the optimum number of clusters may not necessarily lead to the best segmentation results.

Enhancement of OMI aerosol optical depth data assimilation using artificial neural network

A regional chemical transport model assimilated with daily mean satellite and ground-based aerosol optical depth (AOD) observations is used to produce three-dimensional distributions of aerosols throughout Europe for the year 2005. In this paper, the AOD measurements of the Ozone Monitoring Instrument (OMI) are assimilated with Polyphemus model. In order to overcome missing satellite data, a methodology for preprocessing AOD based on neural network (NN) is proposed. The aerosol forecasts involve two-phase process assimilation and then a feedback correction process. During the assimilation phase, the total column AOD is estimated from the model aerosol fields. The main contribution is to adjust model state to improve the agreement between the simulated AOD and satellite retrievals of AOD. The results show that the assimilation of AOD observations significantly improves the forecast for total mass. The errors on aerosol chemical composition are reduced and are sometimes vanished by the assimilation procedure and NN preprocessing, which shows a big contribution to the assimilation process.

Particle Swarm Optimization Based Fast Fuzzy C-Means Clustering for Liver CT Segmentation

A Fast Fuzzy C-Means (FFCM) clustering algorithm, optimized by the Particle Swarm Optimization (PSO) method, referred to as PSOFFCM, has been introduced and applied on liver CT images. Compared to FFCM, the proposed approach leads to higher values in terms of Jaccard Index and Dice Coefficient, and thus, indicating higher similarity with the ground truth provided. Based on ANOVA analysis, PSOFFCM showed better results in terms of Dice Coefficient. It also showed better mean values in terms of Jaccard Index and Dice Coefficient based on the box and whisker plots.

Melanoma detection using fuzzy C-means clustering coupled with mathematical morphology

This paper proposes a Fuzzy C-Means (FCM) based approach designed for melanoma diagnosis. The methodology comprises the traditional data processing architecture, including pre-processing (contrast stretching), main processing (FCM) and post-processing (morphological erosion). The contrast stretching phase has the purpose of stretching the range of pixel intensities of the input image to occupy a larger dynamic range in the output image. This is followed by the FCM algorithm, which automatically divides the data provided by the contrast stretching phase into two clusters: lesion and skin. This process ends with the morphological erosion of the segmented image, where the structuring element is translated over each pixel of the object, so as to overcome typical irregularities between lesion and skin (e.g., irregular boundaries, dark hair covering the lesions, specular reflections, among others). The proposed approach is evaluated in dermatoscopic images of skin cancer, and results show that it is able to produce accurate identification of lesions.

PSilhOuette: Towards an Optimal Number of Clusters Using a Nested Particle Swarm Approach for Liver CT Image Segmentation

This paper proposes a nested particle swarm optimization (PSO) method to find the optimal number of clusters for segmenting a grayscale image. The proposed approach, herein denoted as PSilhOuette, comprises two hierarchically divided PSOs to solve two dependent problems: i) to find the most adequate number of clusters considering the silhouette index as a measure of similarity; and ii) to segment the image using the Fuzzy C-Means (FCM) approach with the number of clusters previously retrieved. Experimental results show that parent particles converge towards maximizing the silhouette value while, at the same time, child particles strive to minimize the FCM objective function.

Liver CT Image Segmentation with an Optimum Threshold Using Measure of Fuzziness

This paper presents a Fuzzy C-Means based image segmentation approach with an optimum threshold using measure of fuzziness. The optimized version, herein denoted as FCM-t, benefits from an optimum threshold, calculated using measure of fuzziness. This allows the revealing of ambiguous pixels, which are eventually assigned to the appropriate clusters by calculating the rounded average cluster values in the ambiguous pixels neighbourhood. The proposed approach showed significantly better results compared to the traditional Fuzzy C-Means, at the cost of some processing power. By benefiting from the optimum threshold approach, one is able to increase the segmentation performance by approximately three times more than with the traditional FCM.

Integration of Neural Network Preprocessing Model for OMI Aerosol Optical Depth Data Assimilation

A regional chemical transport model assimilated with daily mean satellite and ground based Aerosol Optical Depth (AOD) observations is used to produce three dimensional distributions of aerosols throughout Europe for the year 2005. In this paper, the AOD measurements of the Ozone Monitoring Instrument (OMI) are assimilated with Polyphemus model. In order to overcome missing satellite data, a methodology for pre-processing AOD based on Neural Network (NN) is proposed. The aerosol forecasts involve two-phase process assimilation, and then a feedback correction process. During the assimilation phase, the total column AOD is estimated from the model aerosol fields. The model state is then adjusted to improve the agreement between the simulated AOD and satellite retrievals of AOD. The results show that the assimilation of AOD observations significantly improves the forecast for total mass. The errors on aerosol chemical composition are reduced and are sometimes vanished by the assimilation procedure and NN preprocessing, which shows a big contribution to the assimilation process.

A Study of Action Recognition Problems: Dataset and Architectures Perspectives

Action recognition field has recently grown dramatically due to its importance in many applications like smart surveillance, human–computer interaction, assisting aged citizens or web-video search and retrieval. Many research trials have tackled action recognition as an open problem. Different datasets are built to evaluate architectures variations. In this survey, different action recognition datasets are explored to highlight their ability to evaluate different models. In addition, for each dataset, a usage is proposed based on the content and format of data it includes, the number of classes and challenges it covers. On other hand, another exploration for different architectures is drawn showing the contribution of each of them to handle different action recognition problem challenges and the scientific explanation behind their results. An overall of 21 datasets is covered with 13 architectures that are shallow and deep models.

PFastNCA: Parallel Fast Network Component Analysis for Gene Regulatory Network

One of the gene expression data analysis tasks is the Gene regulatory network analysis. Gene regulatory network is concerned in the topological organization of genes interactions. Moreover, the regulatory network is important for understanding the normal cell physiology and pathological phenotypes. However, the main challenge facing gene regulatory network algorithms is the data size. Where, the algorithm runtime is proportional to the data size. This paper presents a parallel algorithm for gene regulatory network (PFastNCA) which is an improved version of FastNCA. PFastNCA enhanced the main core of FastNCA which is the connectivity matrix estimation using a distributed computing model. Where, the work is divided among N processing nodes, PFastNCA is more efficient than FastNCA. It also achieved a better performance and speedup reached 1.91.