Petr Chmelar

SUNAR Surveillance Network Augmented by Retrieval

The paper deals with Surveillance Network Augmented by Retrieval (SUNAR) system – an information retrieval based wide area (video) surveillance system being developed as a free software at FIT, Brno University of Technology. It contains both standard and experimental techniques evaluated by NIST at the AVSS 2009 Multi-Camera Tracking Challenge and SUNAR performed comparably well.

Buffer overflow attacks data acquisition

In this abstract, we investigate the network traffic that may cause the unauthorized control of a computer in the campus network using buffer overflow attacks, the objective of which is to gain the control of privileged programs and computers. We provide statistics of the network traffic in a campus and an eterprise network together with probabilities of a buffer overflow attack to provide attakers the most vulnerable services using low interaction honeypot HoneyD together with a highly interactive shadow honeypot Argos that were used to detect attacks and describe their detection profiles. In this manner, we can collect data to be used for training classifiers to predict and detect even zero day vulnerabilities and malware. Our intension is to acquaint dataset that can identify serious security threats in much higher details, compared to 1999 KDD Cup dataset.

Interactive Visualization of Data-Oriented XML Documents

There is many data stored and exchanged in XML nowadays. However, understanding of the information included in large data-oriented XML instances may not be simple in contrast to the presentation of document-oriented XML instances. Moreover, acquaint the desired knowledge from complex XML data is hard or even impossible.

Towards composable robotics: the R3-COP knowledge-base driven technology platform

The ARTEMIS project R3-COP (Resilient Reasoning Robotic Co-operating Systems) aims at providing European industry with leading-edge innovation that will enable the production of advanced robust and safe cognitive, reasoning autonomous and co-operative robotic systems at reduced cost.This is achieved by cross-sector reusability of building blocks, collected in a knowledge base, within a generic framework and platform with domain-specific instantiations. The R3-COP Framework is targeting at becoming basis for a European RTP (Reference Technology Platform) for robust autonomous systems by embodying methodologies, methods, and tools for safety-critical hard-real-time system development and verification supported by European tool vendors. To enable this, interoperability issues have to be resolved at several levels, including meta-models, models, tool interfaces and component descriptions. The link is established by the knowledge base described in more detail in this paper to allow composition of robotic applications from building blocks, guiding design & development as well as validation & verification (supporting certification in the end). The concept of the knowledge base could be re-used in the planned ARTEMIS Common Reference Technology Platform for critical systems engineering.

Visual Surveillance Metadata Management

The paper deals with a solution for visual surveillance metadata management. Data coming from many cameras is annotated using computer vision units to produce metadata representing moving objects in their states. It is assumed that the data is often uncertain, noisy and some states are missing. The solution consists of the following three layers: (a) data cleaning layer - improves quality of the data by smoothing it and by filling in missing states in short sequences referred to as tracks that represent a composite state of a moving object in a spatiotemporal subspace followed by one camera, (b) Data integration layer - assigns a global identity to tracks that represent the same object, (c) Persistence layer - manages the metadata in a database so that it can be used for online identification and offline querying, analyzing and mining. A Kalman filter technique is used to solve (a) and a classification based on the moving objects state and its visual properties is used in (b). An object model for layer (c) is presented too.

Real-Time Indexing of Complex Data Streams

The paper deals with indexing of a complex type data stream stored in a database. We present a novel indexing schema and framework referred to as ReTIn (Real-Time Indexing), the objective of which is to allow indexing of complex data arriving as a stream to a database with respect to soft real-time constraints met with some level of confidence for the maximum duration of insert and select operations. The idea of ReTIn is a combination of a sequential access to the most recent data and an index-based access to less recent data stored in the database. The collection of statistics makes balancing of indexed and unindexed parts of the database efficient. We have implemented ReTIn using PostgreSQL DBMS and its GIN index. Experimental results presented in the paper demonstrate some properties and advantages of our approach.

Chapter 12 – Advanced Security Network Metrics

In this chapter we propose a method for the extraction of data from network flow and a contextual separation of partial connections, using a set of network metrics that create a signature defining the connection behavior. We begin with defining the input dataset of captured communication and the process of extracting metrics from separated connections. Then we define the set of metrics included in the final behavioral signature. The second part of the chapter describes experiments performed with a state-of-the-art set of network metrics, with comparison to our proposed experimental set. The chapter concludes with the results of our experiments.

VTApi: An Efficient Framework for Computer Vision Data Management and Analytics

VTApi is an open source application programming interface designed to fulfill the needs of specific distributed computer vision data and metadata management and analytic systems and to unify and accelerate their development. It is oriented towards processing and efficient management of image and video data and related metadata for their retrieval, analysis and mining with the special emphasis on their spatio-temporal nature in real-world conditions. VTApi is a free extensible framework based on progressive and scalable open source software as OpenCV for high- performance computer vision and data mining, PostgreSQL for efficient data management, indexing and retrieval extended by similarity search and integrated with geography/spatio-temporal data manipulation.

Clustering for Video Retrieval

The paper deals with an application of clustering we used as one of data reduction methods included in processing huge amount of video data provided for TRECVid evaluations. The problem we solved by means of clustering was to partition the local feature descriptors space so that thousands of partitions represent visual words, which may be effectively employed in video retrieval using classical information retrieval techniques. It has proved that well-known algorithms as K-means do not work well in this task or their computational complexity is too high. Therefore we developed a simple clustering method (referred to as MLD) that partitions the high-dimensional feature space incrementally in one to two database scans. The paper describes the problem of video retrieval and the role of clustering in the process, the MLD method and experiments focused on comparison with other clustering methods in the video retrieval application context.