May 23, 2017,  12:30-13:30,  Location : Zi 2126,  Speaker : Dr. T. Inan (Tolga)

Pose Robust 3D Face Recognition

Over the last decade, three-dimensional facial recognition has been extensively researched. Very good results have been reported for frontal and non-expressive faces. Recent studies have focused on identifying faces with orientation and expression variations. This problem is still being investigated.
We will also refer to our work for three-dimensional face recognition that is robust to orientation. This seminar will contain a short summary from the speaker's previous research experience. 

May 03, 2017,  12:30-13:30,  Location : Zi 2042,  Speaker : Drs. H.C.A. Wienen (Hans)

Elephantophagism, or the hunt for an accident

When conducting a literature review, you want to find as many relevant papers as possible. But how do you select relevant articles from an initial corpus of 1775 articles? How do you eat an elephant?

I present the way in which we surveyed the current state of the art in accident analysis methods, starting out from all 108 databases the University has available and ending in a set of three classes of accident models. How did we make sure that no relevant articles were missed, how did we compare results and which results did we compare?

Furthermore, I'll present some of our findings from the review: which classes of analysis methods are there, in what respects do they differ and what are our plans now that we know this. 

April 12, 2017,  13:00-14:00,  Location : Zi 2042,  Speaker : D.H. Apriyanti (Diah) MSc.

Flower Biometrics: Identification of Orchid Species Using Flower Image

The system to identify orchid species has developed. Although taxonomists usually use the key identification that needs some parts of the plant, this system only needs the orchid's flower image. The system uses semi-automated segmentation process, takes the HSV colour feature also shape features such as Centroid Contour Distance, aspect ratio, roundness, etc from the flower image and then identifies them using k-Nearest Neighbors which is compared with Probabilistic Neural Network and Support Vector Machine. Orchid is a unique flower. It has a part of the flower called lip (labellum) that distinguishes it from other flowers even from other types of orchids. We also proposed to do feature extraction not only on flower region but also on lip (labellum) region. The result shows that our proposed method can increase the accuracy value of the system. 

March 08, 2017,  14:00-15:00,  Location : Zi 2042,  Speaker : P. Normakristagaluh (Pesigrihastamadya (Pesi)) MSc.

Plant Species Classification by Leaf Shape and Vein – A Case Study of the Dillenia Genus

A classification of the architectural features of dicot leaves---i.e., the placement and form of those elements constituting the outward expression of leaf structure, including shape, marginal position and gland position----has been developed as the result of an extensive survey of plants. In this research, image processing methods have been used to classify genus of Dillenia. The feature extraction method is applied as a descriptor that was including a method of fractal dimension, length and width ratio, ratio of perimeter and area, moment invariant and the angle between the primary and secondary venation. After extraction of such features, a selection performed by the method of variance analysis and CFS Subset Evaluator. Identification of Dillenia, which is fundamental application of the classification, is determined by shape and an angle of divergence between primary and secondary vein. Classification method's consists of Naive Bayesian Classifier, FLVQ (Fuzzy Learning Vector Quantization), and BPNN (Back Propagation Neural Network) and ELM (Extreme Learning Machine). 

February 23, 2017,  14:00-15:00,  Location : Zi 2042,  Speaker : Dr. A. Vasenev (Alexandr)

Improving the Robustness of Urban Electricity Networks (IRENE) project

The societal and economic consequences of power outages can be severe, in particular if outages last longer than a few hours. Highly decentralized energy system of future smart cities could use wind and sun energy to mitigate the impact of such outages. The IRENE (Improving the Robustness of Urban Electricity Networks, project concentrated on ways to identify threats to the grid and on how stakeholders can improve grid robustness. This seminar overviews outcomes of the project. 

February 16, 2017,  13:00-14:00,  Location : Zi 2042,  Speaker : Marcos R. S. Borges

Emergencies and Disasters: The application domain for studies on Knowledge Engineering at the GRECO Research Group

GRECO (Knowledge Engineering Group) is a research group that is part of the Graduate Program in Informatics at the Federal University of Rio de Janeiro. The GRECO is formed by seven faculty members, 25 Ph.D. Students and over 35 M.Sc. students. The research topics covered by the group include information and knowledge management, collaboration support technologies, human factors and interfaces, ergonomics, social networks, and cognitive task analysis. Although members of the group have been working with Emergency Management Systems for over 15 years, only recently the group adopted emergencies and disasters as the preferential application domains for its research studies. As a result a number of works are under development using the problems of the various phases of disasters and emergencies cycle as problem motivation for research projects. This talk presents a summary of current projects that focus on the Emergency Response phase. Several concepts guide our research: Context, Knowledge sharing, Cooperation and group Decision Support. It will present the framework that was construct around these concepts.

Short Biography:

DR.MARCOS BORGES is Full Professor in the Computer Science Department at the Federal University of Rio de Janeiro. He earned his doctorate in Computer Science from the University of East Anglia (UK) in 1986. From 1994-1996, he was a visiting research scholar and a member of the Object Technology Lab at Santa Clara U. in California. Dr. Borges also served as Visiting Professor at the Polytechnic U. of Valencia, Spain. He has published over 200 research papers in international conferences and journals, including Decision Support Systems, Computers in Industry, and Expert Systems and Applications. His research interests include Computer Supported Collaborative Work (CSCW), Collective Knowledge and Emergency Management Systems. He is a member of the ISCRAM Board and organized the 2016 edition of ISCRAM Conference in Rio de Janeiro. He is also the PC Co-Chair for the ACM CSCW 2017 Conference. 

February 15, 2017,  13:00-14:00,  Location : Zi 2042,  Speaker : N.H. Lestriandoko (Nova) MSc.

Chip Quality Analysis Using Hough Transform

Collaboration Research between LIPI - ALICE CERN

ALICE (A Large Ion Collider Experiment) is a heavy-ion detector on the Large Hadron Collider(LHC) ring. It is designed to study the physics of strongly interacting matter at extreme energy densities, where a phase of matter called quark-gluon plasma forms. ALICE equipment consists of various detectors with their own functions. Those detectors are Inner Tracking System (ITS), Time Projection Chamber (TPC), Time of Flight Detector (TOF),Muon Spectrometer, Transition Radiation Detector (TRD) and etc. Approximately 600 million times per second, particles collide within the Large Hadron Collider (LHC). Each collision generates particles that often decay in complex ways into even more particles. Electronic circuits record the passage of each particle through a detector as a series of electronic signals, and send the data to the CERN Data Centre (DC) for digital reconstruction.

The sensor chip size measurement and adjusting chip position are parts of quality analysis of detectors in the ITS Project. By the using of Hough Transform, these activities can be done easily. The edge detection, cropping, mean filter, and intensity leveling up were used as image preprocessing. Mexican Hat filter also could be used to support Hough voting, especially to increase the accuracy. Experimental results over several chip images showed the efficiency of proposed method. 

January 25, 2017,  12:30-13:30,  Location : Zi 2042,  Speaker : W.H.A. Alsaqaf (Wasim) MSc.

Services: Engineering Quality Requirements in Large Scale Distributed Agile Environment

Agile software development methods have become increasingly popular in the last years. However, agile methods don't specify explicitly how to deal with the quality requirements. Moreover there is little known about how organizations currently deal with this shortcoming. Based on several case studies this research will investigate real-world large-scale distributed agile projects to understand the challenges agile teams face regarding quality requirements and the approach they are currently using to cope with these challenges. After that a set of good practices will be introduced to explicitly integrate quality requirements in agile processes. Other case studies will be conducted to validate the suggested good practices. 

January 11, 2017,  12:30-13:30,  Location : ZI 2042,  Speaker : E. Haasnoot (Erwin) MSc.

Safety: Trained Behavioural Biometrics

Consider the following hypothetical situation: You have line-up of Christiano Ronaldo (CR) and three uncanny look-a-likes of his. What characteristics can you use to most easily identify who is CR and who is a look-a-like? How about the same situation, but with Lionel Messi and look-a-likes, or with Slash (guitarist of Guns 'n Roses), or with Vincent van Gogh? One method would be to ask the ''genuines'' to respectively play football (both CR and Messi), play guitar and paint. This works because all aforementioned people have trained for years and years are top performers in their respective skills. For a look-a-like to successfully impostor a top performer, he would need to come close to being a top performer himself. And.. there is no way for him to do that other than putting the same years of training in himself.

In other words, skill performance could be used as a biometric authentication measure, providing an interesting link between cognitive psychological theory of learning and biometrics. My research will focus on how take this fact and apply it to a real-world application, Ubiqu's Authenticate system. In this presentation I hope to introduce myself as well as my PhD-sponsor Ubiqu and my research plans. I hope to see you there! 

December 21, 2016,  11:00-12:00,  Location : ZI 2042,  Speaker : D. Zeng (Dan) MSc.

Safety: Towards Resolution Invariant Face Recognition in Uncontrolled Scenarios

Face recognition, as one of the most ubiquitous biometrics technologies, has been extensively studied in recent decades. However, the performance is far from being satisfactory for images captured in uncontrolled scenarios. Face images captured by surveillance cameras usually have poor quality, particularly low resolution (LR), which affects the performance of face recognition seriously. Most existing methods deal with such cross-resolution face recognition problem either by importing the information of high resolution (HR) images to help synthesize HR images from LR images or by applying the discrimination of HR images to help search for a unified feature space.

We develop a novel approach to address the problem of matching a LR face image against a gallery of relatively HR face images. We treat the discrimination information of HR and LR face images equally to boost the performance. The proposed approach learns resolution invariant features with the help of popular deep networks aiming to: (1) classify the identity of both LR and HR face images accurately, and (2) preserve the discriminative information among subjects across different resolutions. We conduct experiments on databases of uncontrolled scenarios, i.e., SCface and COX, and results show that the proposed approach significantly outperforms state-of-the-art methods. 

December 13, 2016,  12:30-13:30,  Location : Zi 2126,  Speaker : Dr. Wang (Chong)

Services: Service-Oriented Business Process Registration for Semantic Interoperation

Abstract: To facilitate business collaboration and interoperation among enterprises, it is critical to discover and reuse appropriate business processes modelled in different languages and stored in different repositories. Therefore, it is a challenge to register them in a unified manner without changing their original representations and semantics. For this purpose, we proposed a reference metamodel for process model registration (PMR) to register selected metadata and semantics of heterogeneous business processes. It can help facilitate the semantic discovery of business processes across enterprises, and promote process interoperation and business collaboration. Furthermore, we introduced a multiple annotation frame called RGPS for the process to be registered based on PMR, and benefit process discovery and reuse at different levels of abstraction. Finally, our approach takes BPEL as an example to illustrate how to specify the mappings from BPEL to PMR when performing PMR-based registration of business processes in a specific modelling language. 

December 08, 2016,  13:30-14:30,  Location : ZI 2042,  Speaker : Dr. S. Mondal (Soumik)

Safety: Continuous User Authentication and Identification: Combination of Security & Forensics

For most computer systems, once the user's identity is verified at login, the system resources are available to that user until he/she locks the computer or session. In fact, the system resources are available to any user during that period. This can lead to session hijacking, in which an attacker targets an open session, e.g. when the genuine user gets a cup of coffee, or leaves to talk to a colleague. To avoid unauthorized use of a computer and its resources is a continuous check of the user's identity extremely important. Continuous Authentication (CA) has built around the biometrics supplied by the user's physical or behavioural characteristics and continuously checks the identity of the user throughout the session. It answers the question ''Is an impostor using the system?''. We do not only focus on CA, but also on Continuous Identification (CI) which can be used for forensic evidence. Here the related question is ''Can the imposter be identified once the CA system detects an imposter using the system?''. This is the first time that the CI issue is addressed in this manner.

We contributed a robust dynamic trust model that can be applied to any CA system irrespective of the biometric modality or computing device. This algorithm is able to make a decision on the genuineness of the user after each and every single action performed by the current user. We found that very few actions were required to detect an imposter user. We applied a novel score boost algorithm and came up with a feature selection technique to improve the results.

We developed an identification technique called Pairwise-User-Coupling (PUC) that can reduce a multi-class classification problem into several two-class classification problems. We applied this technique for CI and achieved high identification accuracy even for weak biometric modalities.​ 

December 06, 2016,  12:30-13:30,  Location : CR 3244,  Speaker : P.M. Singh (Prince) MSc.

Services: An enterprise architecture for context aware logistics

In today's hyper connected world, all business can use real time data for service improvement and better profits. In logistics real time and contextual data can help improve many functions like resource utilization, planning and handling uncertain circumstances. In this seminar, I will present a holistic view over how a logistic company can perform smart logistics. I have used enterprise architecture modeling to represent the process, application and technology of a typical logistic company. I have also proposed the changes necessary to evolve into a smart logistic company. A demo application will also be presented which incorporates the improved business processes for a logistic company. 

November 18, 2016,  14:00-15:00,  Location : ZI 2042,  Speaker : R. Bortolameotti (Riccardo) MSc.

Cybersecurity: Aftermath of a data breach: A secure and reliable log to determine the data leakage

After a system recovery from a successful attack it is needed to determine what data has been leaked. This forensic process is relevant to consider the impact of a data breach. Moreover, it is mandatory according to the law. Existing methods often consider it as all-or-nothing, as well as often fail to secure the evidence against malicious adversaries or use strong assumptions, such as trusted hardware. In some limited cases, data can be processed in the encrypted domain. Although it is computationally expensive, it can ensure that nothing leaks to an attacker. This makes determining the leakage trivial. In other cases, victims can only consider that all data were leaked.

In contrast, our work presents an approach capable of determining the data leakage using a distributed log that securely records all accesses to the data without relying on trusted hardware, and which is not all-or-nothing. We demonstrate our approach to guarantee secure and reliable evidence against even strongest adversaries capable of taking complete control over a machine. For the concrete application of client-server authentication, we show the preciseness of our approach. Also, we demonstrate that it is feasible and can be integrated with existing services.​ 

November 11, 2016,  12:30-13:15,  Location : Zi 2126,  Speaker : Dennis Schroer

Feasibility of End-To-End Encryption using Attribute Based Encryption in Health Care

ABE is an encryption technique which allows to provide end-to-end encryption in situations where data is shared with multiple users,

based on their roles or properties of the data. This is especially useful in situations where data is stored in an online environment, like cloud storage services.

We propose a practical construction using ABE to store and share electronic health records. We perform several experiments to investigate the feasibility and performance of multiple selected ABE schemes in this construction, by means of an example case of a health care application. Our analysis shows that the application of ABE is feasible, and with some minor restrictions a number of the selected schemes are feasible. 

October 26, 2016,  12:30-13:30,  Location : Zi 2126,  Speaker : Dr. F.A. Bukhsh (Faiza)

Services: Availability Incidents in Telecom Domain: A Literature Review

''Communication is considered as important as the supply of water and electricity. The availability of communication infrastructures are threaten by system complexity, accidents, mistakes and attacks on the physical infrastructures. Communication is considered to be completely automatized and the incidents are mostly attributed to the failure of these automated components. Incidents in telecom domain have wide-spread impact as it leads to communication drop-out such as Internet services, mobile services, landlines telecom, and especially the emergency helpline (112) services. Telecom incidents are reported but information is usually confidential, while information can be useful for other providers in-case they face similar incidents. As the information is not available, due to privacy concern, goal of this presentation is to review incident analysis methods and techniques reported in scientific literature.'' 

October 04, 2016,  12:30-13:30,  Location : ZI 2126,  Speaker : T.R. van de Kamp (Tim) MSc.

Cybersecurity: Private Sharing of IOCs and Sightings

Information sharing helps to better protect computer systems against digital threats and known attacks. However, since security information is usually considered sensitive, parties are hesitant to share all their information through public channels. Instead, they only exchange this information with parties with whom they already established trust relationships.

We propose the use of two complementary techniques to allow parties to share information without the need to immediately reveal private information. We consider a cryptographic approach to hide the details of an indicator of compromise so that it can be shared with other parties.

These other parties are still able to detect intrusions with these cryptographic indicators. Additionally, we apply another cryptographic construction to let parties report back their number of sightings to a central party. This central party can aggregate the messages from the various parties to learn the total number of sightings for each indicator, without learning the number of sightings from each individual party.

An evaluation of our open-source proof-of-concept implementations shows that both techniques incur only little overhead, making the techniques prime candidates for practice. 

September 14, 2016,  14:00-15:00,  Location : ZI 2126,  Speaker : Ing. G.J. Laanstra (Geert Jan)

NVIS Field Research: Science or Adventure?

Ionospheric radio wave propagation can help in managing and coordinating efforts after natural disasters, not mentioning communications with remote areas. Near Vertical Incidence Skywave (NVIS) propagation mechanism can connect people located up to 400 km from each other. For this, a highly precise beacon transmitter system have to be carefully designed and implemented. Field research is an important part of this process. This talk describes how a state-of-the-art light-weight, low power, and transportable hybrid antenna-transmitter was developed and tested in Spain and The Netherlands. 

July 06, 2016,  12:30-13:30,  Location : ZI 2042,  Speaker : Joep Peeters (Master student)

Cybersecurity: Fast and Accurate Biometric Verification in the Encrypted Domain

As applications of biometric verification proliferate users become more vulnerable for privacy infringement. Biometric data contains privacy sensitive information like gender, ethnicity and health conditions which should not be shared with third parties during the classification process. Current solutions of secure biometric verification trade security for loss in accuracy and slower processing.

This presentation shows a novel secure biometric verification system which performs classification in the encrypted domain while minimizing the negative effects on accuracy and speed. No private information is shared with the verification service. The system is based on likelihood ratio classification and depends on additive homomorphic ElGamal encryption and a two party secure protocol to perform the verification. 

July 04, 2016,  12:30-13:30,  Location : ZI 2042,  Speaker : M. Caselli (Marco) MSc.

Cybersecurity: Specification Mining for Intrusion Detection in Networked Control Systems

This research discusses a novel approach to specification-based intrusion detection in the field of networked control systems. Our approach reduces the substantial human effort required to deploy a specification-based intrusion detection system by automating the development of its specification rules. We observe that networked control systems often include comprehensive documentation used by operators to manage their infrastructures. Our approach leverages the same documentation to automatically derive the specification rules and continuously monitor network traffic. In this research, we implement this approach for BACnet-based building automation systems and test its effectiveness against two real infrastructures deployed at a major university and a large-scale research center. Our implementation succesfully identifies process control mistakes and potentially dangerous misconfigurations. This confirms the need for an improved monitoring of networked control system infrastructures. 

June 29, 2016,  12:30-13:30,  Location : ZI 2042,  Speaker : MSc J.L. Rebelo Moreira (Joao Luiz)

Services: Developing interoperable early warning systems with an ontology-driven situation-aware framework

In recent years hazard events (e.g. disease outbreaks, tsunamis, earthquakes) have been occurring more often, leading to disaster situations that cause numerous social and economic impacts. An early warning system (EWS) is a system-of-systems for disaster risk reduction (DRR) that provides sensoring, monitoring and forecasting capabilities to cope with disasters. Recently the Sendai framework was established by the main international agencies involved with DRR (leaded by the United Nations) stating the goals to be achieved by 2030. One of the main goals is to achieve the so called multi-hazard EWS by integrating existing EWS and building new interoperable EWS.

In this talk I'll present an overview of my phd research in this context. In particular, I'll talk about how to deal with semantic interoperability, i.e. how to address the challenge of sharing knowledge among all parties that interact with EWS. I'll present the current work regarding an ontology-driven framework we are developing. It is based on the Situation Awareness theory that follows the model-driven engineering paradigm, where implementation is performed with a distributed rule-based engine and complex events processing. Further information refer to:

Moreira, J. L. R., Ferreira Pires, L., Sinderen, M. Van, & Costa, P. D. (2015). Towards ontology-driven situation-aware disaster management. Journal of Applied Ontology, 10(3-4), 339–353. 

June 15, 2016,  11:30-12:30,  Location : ZI2042,  Speaker : Dr. Z. Tan (Zhiyuan)

Cybersecurity: Selecting optimal features for network intrusion detection

Redundant and irrelevant features in data have caused a long-term problem in network traffic classification. These features not only slow down the process of classification but also prevent a classifier from making accurate decisions, especially when coping with big data. We proposed a mutual information based algorithm that analytically selects the optimal feature for classification. This mutual information based feature selection algorithm can handle linearly and nonlinearly dependent data features. Its effectiveness is evaluated in the cases of network intrusion detection. The evaluation results show that our feature selection algorithm contributes more critical features for to achieve better accuracy and lower computational cost compared with the state-of-the-art methods. 

June 01, 2016,  12:30-13:30,  Location : ZI 2042,  Speaker : Wouter Lueks, PhD student (Radboud University)

Cybersecurity: Distributed encryption and applications

Governments increasingly store and process huge quantities of data to combat crime, fraud, and terrorism with the aim of increasing security.

However, the price is a loss of privacy. Fortunately, in some cases, it is possible to build cryptographic systems that achieve the security goals and at the same time protect the privacy of 'the innocent'. One such system is distributed encryption.

Distributed encryption allows observers to record parties that behave suspiciously by creating ciphertext shares of their identities. These shares by themselves give no information about the party to whom they refer. They can only be combined to recover the identity of the recorded party when sufficient shares are available. This system can, for example, be used to find high-way truck-stop robbers without affecting the privacy of regular innocent road users.

In this talk, I will describe our 2014 distributed encryption scheme, including a variant that is much faster for small plaintext domains---like license plates. 

May 25, 2016,  12:30-13:30,  Location : ZI 2042,  Speaker : Prof.Dr. P.H. Hartel (Pieter)

Cyber Security: Block Chain Security Lab

The block chain is a potentially disruptive technology, because it has the potential to erode the role of trusted third parties, such as banks and governments. However, little is known about the criminogenic properties of the block chain and its applications beyond the fact that offenders and terrorists have used Bitcoin. The idea is to set up a lab to research the block chain and its criminogenic properties, and develop crime prevention methods and technology so that the true potential of the block chain and its applications may be realized.