Azrina Aziz

A Geometrical Approach for Age-Invariant Face Recognition

Human faces undergo considerable amounts of variations with aging. While face recognition systems have proven to be sensitive to factors such as illumination and pose, their sensitivity to facial aging effects is yet to be studied. The FRVT (Face Recognition Vendor Test) report estimated a decrease in performance by approximately 5% for each year of age difference. Therefore, the development of age-invariant capability remains an important issue for robust face recognition. This research study proposed a geometrical model based on multiple triangular features for the purpose of handling the challenge of face age variations that affect the process of face recognition. The system is aimed to serve in real time applications where the test images are usually taken in random scales that may not be of the same scale as the probe image, along with orientation, lighting ,illumination, and pose variations. Multiple mathematical equations were developed and used in the process of forming distinct subject clusters. These clusters hold the results of applying the developed mathematical models over the FGNET face aging database. The system was able to achieve a maximum classification accuracy of above 99% when the system was tested over the entire FGNET database.

Age-Invariant Face Recognition Using Triangle Geometric Features

Whilst facial recognition systems are vulnerable to different acquisition conditions, most notably lighting effects and pose variations, their particular level of sensitivity to facial aging effects is yet to be researched. The face recognition vendor test (FRVT) 2012s annual statement estimated deterioration in the performance of face recognition systems due to facial aging. There was about 5% degradation in the accuracies of the face recognition systems for each single year age difference between a test image and a probe image. Consequently, developing an age-invariant platform continues to be a significant requirement for building an effective facial recognition system. The main objective of this work is to address the challenge of facial aging which affects the performance of facial recognition systems. Accordingly, this work presents a geometrical model that is based on extracting a number of triangular facial features. The proposed model comprises a total of six triangular areas connecting and surrounding the main facial features (i.e. eyes, nose and mouth). Furthermore, a set of thirty mathematical relationships are developed and used for building a feature vector for each sample image. The areas and perimeters of the extracted triangular areas are calculated and used as inputs for the developed mathematical relationships. The performance of the system is evaluated over the publicly available face and gesture recognition research network (FG-NET) face aging database. The performance of the system is compared with that of some of the state-of-the-art face recognition methods and state-of-the-art age-invariant face recognition systems. Our proposed system yielded a good performance in term of classification accuracy of more than 94%.

Gasoline-ethanol (Gasohol) fuel blend spray characterization using digital imaging and image processing

This paper tries to characterize the spray of gasoline-ethanol blends using schlieren imaging technique and image processing. Five different gasoline-ethanol fuel blend rates by volume are prepared. The Image of the spray is captured using schlieren imaging technique and image processing is employed to extract macro spray characteristics- spray tip penetration and spray cone angle. Based on the extracted tip penetration and cone angles E55 gasoline-ethanol blend found out to have the largest cone angle and tip penetration among the tested blends.

A Review on Segmentation and Modeling of Cerebral Vasculature for Surgical Planning

Visualization of cerebral blood vessels is vital for stroke diagnosis and surgical planning. A suitable modality for the visualization of blood vessels is very important for the analysis of abnormalities of the cerebrovascular system, as it is the most complex blood circulation system in the human body and vulnerable to bleeding, infection, blood clot, stenosis, and many other forms of damage. Images produced by current imaging modalities are not promising because of noise, artifacts, and the complex structure of cerebral blood vessels. Therefore, there is a requirement for the accurate reconstruction of blood vessels to assist the clinician in making an accurate diagnosis and surgical planning. This paper presents an overall review of modeling techniques that can be classified into the three categories, i.e., image-based modeling, mathematical modeling, and hybrid modeling. Image-based modeling deals directly with medical images and which involves preprocessing, segmentation, feature extraction, and classification. Mathematical modeling exploits existing mathematical laws and equations, an example being an arterial bifurcation, which is assumed to follow a fractal and cube law, and a system of ordinary differential equations are solved to obtain pressure and velocity estimates in a branching network. Whereas, Hybrid modeling incorporates both image-based and mathematical modeling to attempt to produce a more detailed and realistic arterial structure. From the literature review and the analysis of the results, it can be summarized that hybrid models provide a faster and more robust technique, which can significantly help in diagnosis and surgical planning, such as for finding the shortest path for a stenting procedure.

Proposed face recognition system after plastic surgery

Plastic surgery is considered as a challenging research issue in the field of face recognition. Nevertheless, it has yet to be studied from theoretical and experimental perspectives. In this study, the authors proposed a facial recognition system for recognising faces after plastic surgery, which fuses the scores of two feature-based and texture-based algorithms. The feature based algorithm is the image GIST global descriptor and the texture-based algorithm is the local binary pattern (LBP) of silence points. First, the local texture descriptor LBP was applied over a set of key points (silence points) in the face image rather than applying it over the entire face area. This proposed feature set is based on the assumption that only those LBP patterns with certain meaning, such as an edge or corner, will be useful for recognising faces that have undergone plastic surgery. The second set of features was extracted using a global descriptor, which is the GIST descriptor, to obtain a basic and a subordinate level description of the perceptual dimension. The performance of the proposed system surpassed the performance of a number of state-of-the-art face recognition after plastic surgery, with a maximum verification accuracy of more than 91%.

The impact of clustering on the average path length in wireless sensor networks

Clustering algorithms have been widely used in wireless sensor networks for virtual backbone construction. They organize the nodes into smaller groups and form a structured topology allowing more efficient bandwidth usage and battery consumption. As the clustering algorithms are usually used for routing, it is crucial to measure the efficiency of the generated backbone in information transport. Failure to do so will impact the routing performance and reduce the reliability of the system. This paper investigates whether the backbone formed by clustering algorithms is able to preserve the routing paths of the network. This property is evaluated by comparing the performance of several clustering algorithms with respect to the average path length. In order to obtain accurate results, the performance is investigated under different network sizes as well as network densities.

An EEG-Based Cognitive Load Assessment in Multimedia Learning Using Feature Extraction and Partial Directed Coherence

Assessing cognitive load during a learning phase is important, as it assists to understand the complexity of the learning task. It can help in balancing the cognitive load of postlearning and during the actual task. Here, we used electroencephalography (EEG) to assess cognitive load in multimedia learning task. EEG data were collected from 34 human participants at baseline and a multimedia learning state. The analysis was based on feature extraction and partial directed coherence (PDC). Results revealed that the EEG frequency bands and activated brain regions that contribute to cognitive load differed depending on the learning state. We concluded that cognitive load during multimedia learning can be assessed using feature extraction and measures of effective connectivity (PDC).

Dynamic analysis of three degree of freedom (3-DOF) exoskeleton for lower extremities

This paper shows the step by step kinematic and dynamic derivation for three degree of freedom exoskeleton for lower extremities. By using the Denavit Harternberg theorem, the forward and inverse kinematic solutions of the exoskeleton are performed. The Lagrange mathematical statement is utilized which talks about kinetic strength and potential strength amid maneuver. The Anthropometry of the proposed exoskeleton design is computed by using Dempster, Clauser and Chandler method. This outcome can be used in control design for the proposed prototype lower extremities exoskeleton.

Cerebral vessel enhancement using bilateral and Hessian-based filter

Vessel enhancement in magnetic resonance angiography (MRA) is an important preprocessing step for stroke surgical planning and further processing. Bilateral filter has been widely used to reduce noise due to its ability for smoothing an image and preserve the edges. It suffers a drawback of over smoothing that leads to discontinuity of blood vessels when applied to MRA image. Hessian-based filter is known to have the ability of enhancing the vessels and preserving the geometrical structure. This paper presents the vessel enhancement technique which combines bilateral and Hessian-based filters to exploit the advantages of them. The bilateral filtered images show that weighted bilateral filter can reduce more noise when comparing the peak signal-to-noise ratio (PSNR) value. Then, Hessian-based filter is performed on several types of preprocessed images to compare the performance of this technique on different types of images. Our method shows a promising result in suppressing the noise that is enhanced by Hessian-based filter.

Evaluation of rehearsal effects of multimedia content based on EEG using machine learning algorithms

Rehearsal is a common phenomenon of practicing something to make it more resilient in long-term memory. This paper will present the rehearsal effects based on electroencephalography (EEG) recorded data for multimedia contents. Three frequency based features are used to discriminate the three learning states mentioned as L1, L2 and L3 using machine learning algorithms. From these three learning states, L1 is the first learning state whether L2 and L3 are the rehearsal states of L1. The set of spectral features that are used for analysis are based on the intensity weighted mean frequency (IWMF), its bandwidth (IWBW), and spectral power density (PSD). For the analysis, the three brain waves investigated are the alpha waves, theta waves and delta waves. The results of the study show that the alpha waves produce de-synchronization from rest to learning state as compared to other EEG recorded waves. This de-synchronization lead to mental effort imposed by working memory during a learning task. The Alpha wave shows more accuracy in L1 using SVM classifier that is 85% using PSD features, 86% for IWFM and 78.4% using IBWB feature. The results also mention that L3 produces less classifier accuracy value as compared to the L2 and L1 for each of three extracted features. This indicates that L3 requires less mental effort during learning. The findings proved the rehearsal as a good phenomenon of long-term memorized learning.