Conference Papers

MSF1: Computer and Information Systems

Visual Target Tracking Under Varying Illumination Conditions

Buti Al Delail (Khalifa University of Science, Technology and Research, United Arab Emirates (UAE)); Harish Bhaskar (Khalifa University of Science Technology and Research, United Arab Emirates (UAE)); Jamal Zemerly (Khalifa University, United Arab Emirates (UAE)); Mohammed AlMualla (Khalifa University of Science, Technology and Research, United Arab Emirates (UAE))

Abstract

Visual targets tracking under varying illuminations is a challenging problem, mainly when the illumination varies across different regions of the target. This paper surveys the current stateoftheart in visual target tracking under varying illumination conditions, with the main focus on particle filter based trackers. Being one of the main topics in computer vision, much research has contributed to visual target tracking in the past decade. In this context, particle filter based approaches have captured a fair amount of attention. This is because it effectively deals with solving nonlinear and nongaussian probability distribution, as well as its ability to recover from occlusions, thus making it suitable for many tracking problems. However, current particle filtering methods do not perform well when the target appearance changes significantly. By allowing the particle filter to deal with large variations in pose, shape, and illumination, the risk of losing the tracked target is increased.

Multiple Proposals for Sign Language Recognition

Mohamed Hassan, Khaled Assaleh and Tamer Shanableh (American University of Sharjah, United Arab Emirates (UAE))

Abstract

Sign language recognition (SLR) as whole is considered as relatively new area, and Arabic SLR (ArSLR) specifically is quite new as it did not receive much of attention until recent years. This work present a comprehensive comparison between three different classification techniques for continuous ArSLR, namely our Modified knearest neighbor(KNN) and two hidden Markov models (HMM) toolkits. In addition, Two datasets are introduced and will be available for the research community. Both datasets composed of the same 40 Arabic sentences but have been collected using two different acquisition technologies? Polhemus G4 motion trackers and DG5VHand data gloves . We also discuss the advantages and disadvantages of each technology. Each dataset have been classified using each of the the aforementioned classification techniques. This paper and to the best of our knowledge propose the first continuous ArSLR using motion tracker.

Dynamic Textures Based Target Detection for PTZ Cameras

M. Sami Zitouni (Khalifa University, United Arab Emirates (UAE)); Harish Bhaskar (Khalifa University of Science Technology and Research, United Arab Emirates (UAE)); Mohammed AlMualla (Khalifa University of Science, Technology and Research, United Arab Emirates (UAE))

Abstract

In this paper, a temporallyiterative based GMM of DT model for target detection using a moving PTZ camera, is proposed. Camera movement in a PTZ sensor causes motionbased target detection techniques to fail for the periods affected by the scene change, where the whole scene is considered a representation of the target motion. During camera motion, conventional background models remain invalid until the time that the model has adapted and updated its parameters to the newly perceived scene. The proposed model is based on an iterative modeling of spatiotemporal patches representing the visual scene using a Gaussian Mixture Model (GMM) of dynamic textures (DT). The application of the GMMofDT, during the initial iteration is performed to temporally segment the video into clips separating global from local motion. Further, parameters of the GMMofDT model are estimated for each temporal segment and in subsequent iterations applied adaptively to generate the foreground mask.

Study on LBP Variants over Mesh Manifold

Claudio Tortorici and Naoufel Werghi (Khalifa University, United Arab Emirates (UAE)); Stefano Berretti (University of Florence, Italy)

Abstract

Local Binary Pattern is a successful descriptor that showed high performances in 2D and depth images. In this paper the possibility to replicate 2D LBP variants on a real 3D environment is proposed. Extending LBP variants to mesh manifolds permit to take the advantage of their descriptive power applied on a real 3D support, inheriting its strengths. In particular six variants will be presented and explained, analysing the binary pattern generated. Such study allow to optimize the descriptor size and discriminative accuracy.

MSG1: Health and Life Sciences

Relationship between hormonal status, epigenetic markers and response of breast cancer cells to chemotherapy

Wafaa Ramadan, SH, Ekram Saleh, Varsha Menon and Rafat ElAwady (University of Sharjah, United Arab Emirates (UAE))

Abstract

Triple negative breast cancer (TNBC) is more aggressive and poorer prognosis compared to other breast cancer subtypes. So far there are no potential targeted therapy drugs for TNBC patients. Our aim is to investigate the effect of epigenetic modifiers (SAHA and 5azadc) on the expression of ER? and HER2/ERBB2 and on the response of different breast cancer cell lines to Tamoxifen. Our results show differential expression of epigenetic markers in MCF7, SkBr3, BT549 and MDAMB231 cells. Therefore, treatment with epigenetic modifiers showed different effects on the expression level of ER? and HER2/ERBB2 in the four cell lines. In addition, the sensitivity of these four cell lines to Tamoxifen upon combination with epigenetic modifiers was increased. Moreover, SAHA and 5azadc have a role in inducing apoptosis through decreasing the expression of cMYC, noncleaved Bid and Bclxl while increasing the expression of p53. The results of the present study indicate that modification of epigenetic status of breast cancer cells specially the triple negative ones enhances their response to hormonal therapy through upregulation of ER? and induction of apoptosis.

The Molecular and Cellular Mechanisms of Desbuquois Dysplasia type II

Nesreen Al Jezawi (UAE University, United Arab Emirates (UAE)); Lihadh AlGazali and Bassam Ali (United Arab Emirates University, United Arab Emirates (UAE))

Abstract

Desbuquois syndrome is a rare heterogeneous skeletal dysplasia characterized by shortlimbed dwarfism, dysmorphic facial features and severe joint laxity. Two types have been recognized based on the presence of distinctive carpal and phalangeal features. Mutations in CANT1 gene have been found to be responsible for type I and lately, for Kim type. In addition, a number of Desbuquois dysplasia type II patients have been attributed to mutations in xylosyltransferase1, encoded by the XYLT1 gene, an enzyme essential for the biosynthesis of proteoglycans. We report here a patient with Desbuquois dysplasia II with a novel homozygous duplication of a single nucleotide in XYLT1 gene (c.2169dupA). The variant is predicted to result in a frameshift and stop codon (p.Val724Serfs*10) within the xylosyltransferase catalytic domain. Confocal analysis of the current and the previously reported missense mutations (c.1441C>T, p.Arg481Trp and c.1792C>T, p.Arg598Cys), revealed aberrant subcellular localization of the mutant proteins, suggesting endoplasmic reticulum retention.

The Role of ShcD and Ret interaction in neuroblastoma survival and migration

Zeanap Mohamed (University Of Sharjah, United Arab Emirates (UAE))

Abstract

Preliminary screening data showed that ShcD adaptor protein associates with the protooncogene Ret receptor tyrosine kinase. In the present study, we aimed to investigate the molecular interaction between ShcD and Ret in human neuroblastoma and studying the functional impact of this interaction. We were able to show that ShcD immunoprecipitated with Ret from SKNAS neuroblastoma cells lysate upon GDNF treatment. This result was validated by ShcDRet colocalization via employing fluorescence microscope. ShcDRet association was shown to have unexpectedly negative impact on various downstream signaling pathways such as AKT, PKC and ERK pathways. Interestingly, ShcDRet association reduced cell viability and migration in SKNAS cell line. However, ShcD was shown previously to trigger melanoma cell migration and tumorigenesis? our data showed an opposite role for ShcD in neuroblastoma SKNAS cells via its association with Ret. Conclusively, ShcD acts as a switch molecule in bringing a contrast biological responses, which needs further determination.

DoxorubicinEncapsulated, EstroneAppended Liposomes Triggered by Ultrasound for the Treatment of Breast Cancer

Pierre Kawak, Vinod Paul, Paul Kawak, Rita Kassermally, Fatme Lahib, Rute Vitor and Mohammad AlSayah (American University of Sharjah, United Arab Emirates (UAE)); Ghaleb Husseini (AUS, United Arab Emirates (UAE))

Abstract

Doxorubicin, like most chemotherapeutic drugs, causes unwanted side effects by interacting with healthy cells. To avoid this and to increase drug delivery efficiency, smart Drug Delivery Systems (DDS) are used. The DDS studied incorporated estrone ligands on its liposomal surface to interact with breast cancer cells and was triggered to release the drug with focused ultrasound. The release properties, sizes and stability of the DDS were studied. The liposome was characterized as a large unilamellar vesicle. The release rate upon insonation increased with increasing power density for constant frequency. Ultrasound is a promising factor for its biocompatibility and noninvasive nature.

MSH1: Arts, Humanities and Social Sciences

Constructive Alignment Challenges

Moza Salim Al Naimi (Masdar Institute of Science and Technology, United Arab Emirates (UAE))

Abstract

Academic sector is in needs of a process that provides a framework for adjusting teaching and assessment to address the achievement of class, department and university outcomes (Biggs, 2014), but the challenges for such a process are unstudied empirically. Should Constructive Alignment (CA) become the standard alignment method in our modern education system? What are causes and consequences (challenges) of such a move, if it is considered? In specific, this study tests wither finance and attendant staff workloads, teacher resistance to change, promotions base on research productivity rather than teaching quality and leadership at all departmental, faculty and institutional levels are the keys challenges for Constructive Alignment implementation. This study will identify for academic sector which of the above factors are representing a real challenges to CA implementation in education system.

Is the Upper Half of Linguistic Stimuli Always More Informative than the Lower? Evidence from English and Arabic

Maryam AlJassmi (Zayed University & Psychology, United Arab Emirates (UAE))

Abstract

Previous studies have reported that when presenting just the upper or lower halves of words, the upper half is more informative than the lower half in a range of different languages. However, despite the widespread usage of Arabic, the informativeness of the upper and lower halves of Arabic words have never been reported. Accordingly, the aim of this study was to investigate the informativeness of the upper and lower halves of Arabic words, nonwords and single letters. ArabicEnglish bilinguals participated in a twoalternative forced choice procedure in which Arabic and English words, nonwords and single letters were displayed in three different formats (complete, upper half, lower half). Results revealed an upper advantage for all English stimuli but a lower advantage for all Arabic stimuli. The magnitude of this upperlower difference was greatest for words in both languages indicating that these effects reflect components specific to word recognition.

Useful
Links
Educating the individual is this country's most valuable investment. It represents the foundation for progress and development. -H.H. Sheikh Khalifa Bin Zayed Al Nahyan
Education is a top national priority, and that investment in human is the real investment to which we aspire. -H.H. Sheikh Mohammed Bin Zayed Al Nahyan

Login For Uae GSRC

Forgot your password reset here

If you do not have an EDAS login Register Here

Online Submission is currentlyclosed.