Showing papers by "Timo Hämäläinen published in 2014"

Efficient Mode Decision Schemes for HEVC Inter Prediction

Abstract: The emerging High Efficiency Video Coding (HEVC) standard reduces the bit rate by almost 40% over the preceding state-of-the-art Advanced Video Coding (AVC) standard with the same objective quality but at about 40% encoding complexity overhead. The main reason for HEVC complexity is inter prediction that accounts for 60%-70% of the whole encoding time. This paper analyzes the rate-distortion-complexity characteristics of the HEVC inter prediction as a function of different block partition structures and puts the analysis results into practice by developing optimized mode decision schemes for the HEVC encoder. The HEVC inter prediction involves three different partition modes: square motion partition, symmetric motion partition (SMP), and asymmetric motion partition (AMP) out of which the decision of SMPs and AMPs are optimized in this paper. The key optimization techniques behind the proposed schemes are: 1) a conditional evaluation of the SMP modes; 2) range limitations primarily in the SMP sizes and secondarily in the AMP sizes; and 3) a selection of the SMP and AMP ranges as a function of the quantization parameter. These three techniques can be seamlessly incorporated in the existing control structures of the HEVC reference encoder without limiting its potential parallelization, hardware acceleration, or speed-up with other existing encoder optimizations. Our experiments show that the proposed schemes are able to cut the average complexity of the HEVC reference encoder by 31%-51% at a cost of 0.2%-1.3% bit rate increase under the random access coding configuration. The respective values under the low-delay B coding configuration are 32%-50% and 0.3%-1.3%.

Analysis of HTTP Requests for Anomaly Detection of Web Attacks

Abstract: Attacks against web servers and web-based applications remain a serious global network security threat. Attackers are able to compromise web services, collect confidential information from web data bases, interrupt or completely paralyze web servers. In this study, we consider the analysis of HTTP logs for the detection of network intrusions. First, a training set of HTTP requests which does not contain any attacks is analyzed. When all relevant information has been extracted from the logs, several clustering and anomaly detection algorithms are employed to describe the model of normal users behavior. This model is then used to detect network attacks as deviations from the norms in an online mode. The simulation results presented show that, compared to other data mining algorithms, the method results in a higher accuracy rate.

Detection of zero-day malware based on the analysis of opcode sequences

Abstract: Today, rapid growth in the amount of malicious software is causing a serious global security threat. Unfortunately, widespread signature-based malware detection mechanisms are not able to deal with constantly appearing new types of malware and variants of existing ones, until an instance of this malware has damaged several computers or networks. In this research, we apply an anomaly detection approach which can cope with the problem of new malware detection. First, executable files are analyzed in order to extract operation code sequences and then n-gram models are employed to discover essential features from these sequences. A clustering algorithm based on the iterative usage of support vector machines and support vector data descriptions is applied to analyze feature vectors obtained and to build a benign software behavior model. Finally, this model is used to detect malicious executables within new files. The scheme proposed allows one to detect malware unseen previously. The simulation results presented show that the method results in a higher accuracy rate than that of the existing analogues.

Artificial Immune System Based Intrusion Detection: Innate Immunity using an Unsupervised Learning Approach

Abstract: This paper presents an intrusion detection system architecture based on the artificial immune system concept. In this architecture, an innate immune mechanism through unsupervised machine learning methods is proposed to primarily categorize network traffic to “self” and “non-self” as normal and suspicious profiles respectively. Unsupervised machine learning techniques formulate the invisible structure of unlabeled data without any prior knowledge. The novelty of this work is utilization of these methods in order to provide online and real-time training for the adaptive immune system within the artificial immune system. Different methods for unsupervised machine learning are investigated and DBSCAN (density-based spatial clustering of applications with noise) is selected to be utilized in this architecture. The adaptive immune system in our proposed architecture also takes advantage of the distributed structure, which has shown better self-improvement rate compare to centralized mode and provides primary and secondary immune response for unknown anomalies and zero-day attacks. The experimental results of proposed architecture is presented and discussed.

An Efficient Network Log Anomaly Detection System Using Random Projection Dimensionality Reduction

Abstract: Network traffic is increasing all the time and network services are becoming more complex and vulnerable. To protect these networks, intrusion detection systems are used. Signature-based intrusion detection cannot find previously unknown attacks, which is why anomaly detection is needed. However, many new systems are slow and complicated. We propose a log anomaly detection framework which aims to facilitate quick anomaly detection and also provide visualizations of the network traffic structure. The system preprocesses network logs into a numerical data matrix, reduces the dimensionality of this matrix using random projection and uses Mahalanobis distance to find outliers and calculate an anomaly score for each data point. Log lines that are too different are flagged as anomalies. The system is tested with real-world network data, and actual intrusion attempts are found. In addition, visualizations are created to represent the structure of the network data. We also perform computational time evaluation to ensure the performance is feasible. The system is fast, finds intrusion attempts and does not need clean training data.

Always Best Connected Heterogeneous Network Concept

Abstract: The concept of always-best-connected (ABC) refers to a person's ability to connect and use services with devices and access technologies that best suit to his or her needs. One of the recent standards to cover this concept is the 3rd Generation Partnership Project evolved packed system (EPS). It offers, an operator-friendly way to manage ABC connectivity in a heterogeneous network environment. However, the existing mobile devices are not supporting all of the requirements of EPS. At the same time, the standard is going through evolution where key elements such as mobility management protocols and procedures are still being developed. This research reviews and addresses constrains to the ABC concept through implementations in Android environment. The research shows that off-the-shelf devices are lacking the needed handover performance for real-time communication in a heterogeneous network. With the help of slightly modified device environments for simultaneous connections over different network access technologies, the results were approaching near real-time communication capability (delay around 100---200 ms depending of selected approach). The trade-offs in terms of energy consumption were measured. The discussion then leads to multihomed devices for really tapping the potential of seamless handovers with the ABC concept.

On optimal relay placement for improved performance in non-coverage limited scenarios

Abstract: Low power nodes have been a hot topic in research, standardization, and industry communities, which is typically considered under an umbrella term called heterogeneous networking. In this paper we look at the problem of deploying optimally low power nodes in the context of relay networking, when an operator connects low power nodes (or small cells) via the wireless backhaul that uses the same spectrum and the same wireless access technology. We present an analytical model that can calculate optimal coordinates for low power nodes based on the input parameters, such as preferred number of nodes, their transmission power, parameters of the environment etc. The analytical calculations are complemented by extensive dynamic system level simulations, by means of which we analyze overall system performance for the obtained coordinates. We also show that even relatively marginal deviations from optimal coordinates can lead to worse system performance.

Experiences from System-on-Chip design courses

Abstract: This paper presents our experiences in teaching System-on-Chip design at Tampere University of Technology. We concentrate on teaching practices and the exercise work which teaches intellectual property (IP) reuse, platform-based design, and HW/SW co-design. Students implement a video encoder on FPGA. It has proved good that teachers split the large exercise into mandatory weekly tasks and bonus points are awarded for good work. Automated testbenches and startup examples were very useful. Students consider the course motivating, even if they underestimate the required time and effort. In the future, more effort is needed in finalizing the reusable IP blocks and whole project works for easier reuse as well as setting up a common repository.

Variable group selection based on regression trees: Paper machine case study

Abstract: This paper presents a methodology for selecting best groups of predictor variables based on regression trees. Test results of the developed methodology applied to industrial pilot paper machine data are presented. Specifically, the results list process variable groups, which are more valuable in predicting paper quality variables. The benefit of paper quality prediction based on process variables is the timely reaction to changes happening during production process and, thus, the reduced operational costs. The proposed regression trees based group variable ranking methodology shows stable results on both data sets used in this study.

Gamification of System-on-Chip design

Abstract: This paper presents drastic new concepts for SoC design tools. The motivation is ever increasing productivity gap that is caused by complexity of the design process with current methods. Plain code, schematic block diagrams and various graph-tree-table descriptions have been there for tens of years. Unfortunately, designers have problems in understanding the big picture of the design and quickly finding the interesting points. As an example, capturing register/memory definitions are simply getting impossible when different operation points of the system change them at runtime. We envision gamification as a new SoC design paradigm. It includes 3D visualization and semantic zoom to capture the design content, and gameplay for the design process. As the result, SoC design becomes engaging and users can easier focus on making the essential design decisions. A demo game is under construction and will be presented in forthcoming publications.

Comparative study of 8 and 10-bit HEVC encoders

Abstract: This paper compares the rate-distortion-complexity (RDC) characteristics of the HEVC Main 10 Profile (M10P) and Main Profile (MP) encoders. The evaluations are performed with HEVC reference encoder (HM) whose M10P and MP are benchmarked with different resolutions, frame rates, and bit depths. The reported RD results are based on bit rate differences for equal PSNR whereas complexities have been profiled with Intel VTune on Intel Core 2 processor. With our 10-bit 4K 120 fps test set, the average bit rate decrements of M10P over MP are 5.8%, 11.6%, and 12.3% in the all-intra (AI), random access (RA), and low-delay B (LB) configurations, respectively. Decreasing the bit depth of this test set to 8 lowers the RD gain of Ml OP only slightly to 5.4% (AI), 11.4% (RA), and 12.1% (LB). The similar trend continues in all our tests even though the RD gain of M10P is decreased over MP with lower resolutions and frame rates. M10P introduces no computational overhead in HM, but it is anticipated to increase complexity and double the memory usage in practical encoders. Hence, the 10-bit HEVC encoding with 8-bit input video is the most recommended option if computation and memory resources are adequate for it.

WOKE: A novel workflow model editor

Abstract: This paper presents a new work-in-progress tool, WOKE, to capture and present the workflow model in a novel way. A workflow model consists of linked steps to describe how actual work is done in a company and usually used for analysis and simulation. The complexity of modern System-on-Chip designs motivates applying such enterprise oriented models to the embedded design process domain as well. For example, UML Activity diagrams, YAWL, BPMN2.0, and flowchart descriptions have been used to capture the product design processes. They support very complex models, but that also leads to several different views that complicate the usability and comprehension. The key idea in WOKE is a very rapid creation and editing of the workflow by textual input, separate real-time graph visualization, comparison, and automatic summaries. User can select which details are visualized. Layout is automated and currently it uses a single column for all tasks and transitions. WOKE is implemented in C++/QT5 and is available as an open source tool. A proof-of-concept case study is shown for an FPGA-based SoC design.

Implementation of Multicore communications API

Abstract: This paper presents an implementation of Multicore Communications API (MCAPI), with focus on portability, stability, and simplicity of the design. The main motivation for the implementation is instability of other publicly available implementations. The developed implementation utilizes POSIX message queues, that is an easily portable interface and readily compatible with MCAPI. The performance was measured as latency and transfer rate of the API. The measurement platforms were a x86-64 PC and a development board featuring an ARM processor. A MCAPI implementation was used as reference for comparison. PMQ-MCAPI is much more stable and easily usable than other MCAPI implementations publicly available for PC. When transfer size was between 1–8 KiB, latency of transfers between cores was between 9–15 µs and transfer rate 500–5000 MBps. This translates to 27 000–45 000 cycles and 0.16–1.67 bytes per cycle. CPU and especially performance of its cache were concluded as the most important factors contributing to the performance. In comparison to the reference, latency of the implementation was 1/8 at best, while transfer rate was up to 35x.

On Detecting the Shape of an Unknown Object in an Electric Field

Abstract: The problem discussed in this paper is detecting the shape of an unknown object in a 2-dimensional static electric field. For simplicity, the problem is defined in a partially rectangular domain, where on a part of the boundary the potential and/or its normal derivative are known. On the other part of the boundary the boundary curve is unknown, and this curve is to be determined. The unknown part of the boundary curve describes the shape of the unknown object.

Wind Turbine Sensor Data Analysis and Production Forecast

Abstract: In this paper we used wind power and meteorological data provided by a Finnish energy company and the Finnish Meteorological Institute as the research material. The study determined the most important factors which had influence to the effectiveness of the wind turbine power production. This was done by using the physical power function with statistical data analysis. Wind speed was found to be the most significant factor for the model. This was due to the fact that wind speed was the only variable which affect was exponential. Another significant factor when it comes to creating a forecast model was temperature. The affect wasn't as powerful as with wind speed but still notable. These observations were also confirmed in statistical interpretation. A tailored forecasting model was formed for our target wind turbine on the basis of these factors: suitable modelling for necessary meteorological factors was executed and the coefficient factor was calculated. The results and especially the forecast model was seen significant and would be used in creation of a production forecast program's first version for the energy company in question.