Hani Mahdi

Agent based correlation model for intrusion detection alerts

Alert correlation is a promising technique in intrusion detection. It analyzes the alerts from one or more intrusion detection system and provides a compact summarized report and high-level view of attempted intrusions which highly improves security effectiveness. Correlation component is a procedure which aggregates alerts according to certain criteria. The aggregated alerts could have common features or represent steps of pre-defined scenario attacks. Correlation approaches composed of a single component or a comprehensive set of components. The effectiveness of a component depends heavily on the nature of the dataset analyzed. The order of correlation component will affect the correlation process performance. Moreover not all components should be used for different dataset. This paper presents an agent-based alert correlation model. Learning agent learns the nature of dataset within a network then guides the whole correlation process and components in such a suitable way of which components could be used and in which order. The model improves the performance of correlation process by selecting the proper components to be used. This model assures minimum alerts to be processed on each component depending on the dataset and minimum time for correlation process.

Recognizing human activity in mobile crowdsensing environment using optimized k-NN algorithm

Abstract Mobile crowdsensing is a recent model in which a group of mobile users uses their smart devices such as smartphones or wearable devices to cooperatively perform a large-scale sensing task. In this paper, a novel model will be introduced for recognizing/classifying human activities that were collected from sensor units on the chest, legs, and arms. The proposed model employed the k-Nearest Neighbor (k-NN) classifier which is one of the most common classifiers. k-NN has only one parameter, k, to determine the number of selected nearest neighbors to the test or unknown samples for predicting the class labels of the unknown samples. Searching for the value of k which has a great impact on the classification performance is difficult especially with high dimensional data. This paper employs the Particle Swarm Optimization (PSO) algorithm to search for the optimal value of the k parameter in the k-NN Classifier. This paper shows first experimentally how the PSO in the proposed algorithm searches for the optimal value of k parameter to reduce the misclassification rate of the k-NN classifier. Then, in the second experiment, ten standard datasets are utilized to benchmark the performance of the proposed algorithm. For verification, the results of the PSO-kNN algorithm are compared with two well-known algorithms: Genetic Algorithm (GA) and Ant Bee Colony Optimization (ABCO). In the third experiment, the proposed PSO-kNN algorithm was employed for recognizing human activities. The experimental results proved that the PSO-kNN algorithm is able to find the optimal or near optimal value(s) of the k parameter which enhances the accuracy of k-NN classifier. The results also demonstrated lower error rates compared when GA and ABCO algorithms.

Variational approach for segmentation of lung nodules

Lung nodules from low dose CT (LDCT) scans may be used for early detection of lung cancer. However, these nodules vary in size, shape, texture, location, and may suffer from occlusion within the tissue. This paper presents an approach for segmentation of lung nodules detected by a prior step. First, regions around the detected nodules are segmented; using automatic seed point placement levels sets. The outline of the nodule region is further improved using the curvature characteristics of the segmentation boundary. We illustrate the effectiveness of this method for automatic segmentation of the Juxta-pleural nodules.

Face sketch recognition using local invariant features

Face sketch recognition is one of the recent biometrics, which is used to identify criminals. In this paper, a proposed model is used to identify face sketch images based on local invariant features. In this model, two local invariant feature extraction methods, namely, Scale Invariant Feature Transform (SIFT) and Local Binary Patterns (LBP) are used to extract local features from photos and sketches. Minimum distance and Support Vector Machine (SVM) classifiers are used to match the features of an unknown sketch with photos. Due to high dimensional features, Direct Linear Discriminant Analysis (Direct-LDA) is used. CHUK face sketch database images is used in our experiments. The experimental results show that SIFT method is robust and it extracts discriminative features than LBP. Moreover, different parameters of SIFT and LBP are discussed and tuned to extract robust and discriminative features.

Feature extraction based on the embedded zero-tree DCT for face recognition

In this paper, a feature extraction method based on the embedded zero-tree of the discrete cosine transform (DCT) is proposed and named EZDCT. In the proposed EZDCT, the DCT is first performed and the embedded zero-tree is then utilized to exploit the relation between the DCT coefficients as in the embedded zero-tree wavelet (EZW). Therefore a subset of the most effective coefficients can be easily selected. The EZDCT and EZW feature extraction methods are compared with the well known methods such as PCA and DCT with the zigzag scan. Experiments are conducted on the standard ORL and FERET databases Experimental results show that the EZDCT achieves a significant improvement in recognition rate with high dimensionality reduction of the feature vector as compared with the other methods. To further improve the recognition performance, the EZDCT and EZW are combined.

A simple evaluation of face detection algorithms using unpublished static images

In this paper, we document the face detection competition that we have organized in conjunction with the ISDA 2010 conference. The objective was to compare different face detection engines performance on new unpublished datasets. We believe researchers can benefit from this competition by identifying strong and weak areas in their algorithms relative to others. We have also identified, based on the results, the common areas of improvement necessary for real life scenario applications.

MASCE: A Multi-Agent System for Collaborative E-Learning

E-learning has become one of the most popular teaching methods in recent years. One of its modes is the blended learning where learners can read teaching materials asynchronously from a teaching website and collaborate with their peers, while providing for necessary face-to-face explanation, discussion, and physical operation in the classroom. In the computational intelligence field, the intelligent agent paradigm gained a tremendous interest in many application domains over the last two decades . This research project paper focuses on the use of intelligent agents in the sphere of e-learning education with the help of collaborative learning. Intelligent agents - the so called e-assistants or helper programs - can sit inside a computer and make the learning in e- learning happen dynamically to suit the need of the user. They can trap the users likes and dislikes in various areas, the level of knowledge and the learning style and accordingly recommend the best matching helpers for collaboration. The paper introduces a multi-agent system for collaborative e-learning (MASCE). MASCE is to assist teaching and learning process and also to encourage collaborative learning among peers. This system shall be used in a blended learning environment as a supplement to the face-to-face lecture where students can use the system in the lab or from home after attending the traditional lecture in the faculty. The objective is to incorporate the intelligence of the multi- agent system (MAS) in a way that enables it to actively and intelligently support the educational processes, where multiple agents can interact to exchange information so that students may collaborate on how best to gain knowledge.

Face Sketch Synthesis and Recognition Based on Linear Regression Transformation and Multi-Classifier Technique

Biometric technique becomes essential to identify individuals in different applications. Face sketch is one of the biometric methods, which are used to identify criminals. In this paper, a face sketch synthesis and recognition model is proposed. In this model, the photo images are transformed to pseudo-sketch images using linear regression technique. Moreover, Gabor filters are used to extract the features from three scales of the images. For each scale, a face sketch image is matched with face pseudo-sketches instead of the original photos to identify an unknown individual. Minimum distance classifier is used to match the sketches with pseudo-sketches in each scale. Further, a classification level fusion is used to combine the outputs of the classifiers at three scales namely, decision, rank, and score level fusion. CHUK database images is used in our experiments. The experimental results show that the proposed model is superior to other existed models in terms of identification accuracy. Moreover, matching sketch images with pseudo-sketches achieved accuracy better than matching sketch images with the original photo images. The proposed model achieved a recognition rate ranged from 82.95 to 94.32 % using single scale, while the accuracy increased to 94.32, 96.6 and 97.7 % when the decision, rank, and score level fusion, respectively, are used.

A Kalman-based encoder for electrical stimulation modulation in a thalamic network model

Restoring vision is no longer impossible as a result of recent advances in neural interfaces. Successful demonstrations of retinal implants motivate the development of more effective visual prostheses. The thalamic Lateral Geniculate Nucleus (LGN) is one potential deep-brain interfacing site for visual prostheses. A main challenge in developing thalamic as well as other visual prostheses is optimizing the parameters of electrical stimulation. This paper introduces a Kalman-based optimal encoder whose function is to determine the optimal electrical stimulation parameters required to induce a certain visual sensation. The performance of the proposed approach is demonstrated using a probabilistic model of LGN neurons. Results demonstrate a significant similarity between neuronal responses obtained using electrical stimulation and the responses obtained using the corresponding visual stimuli with a mean correlation of 0.62 (P <; 0.01, n = 54). These results indicate the efficacy of the proposed optimal encoder in driving LGN neurons to induce visual sensations.

Gender Classification Using Mixture of Experts from Low Resolution Facial Images

In this study, we propose a novel two-stages mixture of experts scheme estimating gender from facial images. The first stage combines a couple of complementary gender classifiers with a third arbiter in case of decision discrepancy. Experimentally, we have verified the common thinking that one appearance-based (Haar-features cascade) classifier with another shape-based (landmarks positions metrology with SVM) classifier form a complementary couple. Subsequently, the second stage in our scheme is a Bayesian framework that is activated only when the arbiter cannot take a confident decision. We demonstrate that the proposed scheme is capable of classifying gender reliably from faces as small as 16x16 thumbnails on benchmark databases, achieving 95% gender recognition on FERET database, and 91.5% on the Labeled Faces in the Wild dataset.